WO2025178982A1

WO2025178982A1 - Methods of powering an implant when misaligned and transmitting data from the implant

Info

Publication number: WO2025178982A1
Application number: PCT/US2025/016517
Authority: WO
Inventors: Kaiyuan Yang; Wei Wang; Zhanghao YU
Original assignee: William Marsh Rice University
Current assignee: William Marsh Rice University
Priority date: 2024-02-19
Filing date: 2025-02-19
Publication date: 2025-08-28
Anticipated expiration: 2026-08-19

Abstract

A system includes an implant and external transceiver, an alignment between them is unknown. A first implant transducer is configured to induce a first field. Each external transducer is configured to receive the first field and convert it to a first electrical signal. An external controller is configured to receive the first electrical signal, determine a received amplitude and phase of the first electrical signal, and determine a first transmitted amplitude and phase of a first electrical driving signal and a second transmitted amplitude and phase of a second electrical driving signal using the received amplitude and phase. One external transducer is configured to induce, using the first electrical driving signal, a second field. Another external transducer is configured to induce, using the second electrical driving signal, a third field. A second implant transducer is configured to receive the second and third fields to power the implant.

Description

METHODS OF POWERING AN IMPLANT WHEN MISALIGNED AND TRANSMITTING DATA FROM THE IMPLANT

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR

DEVELOPMENT

[0001] This invention was made with government support under Grant Numbers 2146476 and ECCS-2023849 awarded by the National Science Foundation. The government has certain rights in the invention.

BACKGROUND

[0002] Wireless, miniature, and minimally invasive bioelectronic implants enable a wide range of applications in healthcare, medicine, and scientific research. In particular, bioelectronic implants promise revolutionary therapies for cardiovascular and neurological disorders.

[0003] Wirelessly powered bioelectric implants eliminate the need for bulky batteries and wires. However, the battery -free implant presents two major challenges. First, despite successful demonstrations of wireless power transfer (WPT) in bioelectronic implants with animal models, the robustness and efficiency of WPT degrade significantly under bioelectronic implant misalignment relative to an external transceiver. Unavoidable misalignment may be incurred by body movements, respiration, heart beating, and limited control of implant orientation intraoperatively. Second, achieving low-power and high- quality uplink communication between the wireless bioelectronic implant and external transceiver efficiently and safely remains challenging.

SUMMARY

[0004] This summary is provided to introduce a selection of concepts that are further described below in the detailed description and examples. This summary is not intended to identify key or essential features of the claimed subject matter, nor is it intended to be used as an aid in limiting the scope of the claimed subject matter.

[0005] In general, in one aspect, embodiments relate to a system. The system includes an implant and an external transceiver disposed up to a predetermined distance away from the implant. An alignment of the implant relative to the external transceiver is unknown. The implant includes a first implant transducer and a second implant transducer. The first implant transducer is configured to induce a first field. The external transceiver includes an external transmitter transducer array, which includes at least two external transducers, and an external controller electrically coupled to the external transmitter transducer array. Each of the at least two external transducers is configured to receive the first field and convert the first field to a first electrical signal. The external controller is configured to, for each of the at least two external transducers, receive the first electrical signal and determine a received amplitude of the first electrical signal and a received phase of the first electrical signal. The external controller is further configured to determine a first transmitted amplitude of a first electrical driving signal and a first transmitted phase of the first electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers and determine a second transmitted amplitude of a second electrical driving signal and a second transmitted phase of the second electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers. The first transmitted phase and the second transmitted phase are a same phase or an opposing phase. A first external transducer among the at least two external transducers is configured to induce, using the first electrical driving signal, a second field. A second external transducer among the at least two external transducers is configured to induce, using the second electrical driving signal, a third field. The second implant transducer is aligned relative to the first implant transducer and configured to receive the second field and the third field and convert the second field and the third field to a second electrical signal. The second electrical signal is configured to power the implant.

[0006] In general, in another aspect, embodiments relate to a method. The method includes inducing, using a first implant transducer of an implant, a first field. The implant is disposed up to a predetermined distance away from an external transceiver. An alignment of the implant relative to the external transceiver is unknown. The method further includes receiving, by an external transmitter transducer array of the external transceiver, the first field and, for each of the at least two external transducers among the external transmitter transducer array, converting the first field to a first electrical signal and determining a received amplitude of the first electrical signal and a received phase of the first electrical signal. The method still further includes determining a first transmitted amplitude of a first electrical driving signal and a first transmitted phase of the first electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers and determining a second transmitted amplitude of a second electrical driving signal and a second transmitted phase of the second electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers. The first transmitted phase and the second transmitted phase are a same phase or an opposing phase. The method also includes inducing a second field by applying the first transmitted amplitude at the first transmitted phase to a first external transducer among the at least two external transducers and inducing a third field by applying the second transmitted amplitude at the second transmitted phase to a second external transducer among the at least two external transducers. The method also includes receiving, by a second implant transducer of the implant, the second field and the third field, converting, using the second implant transducer, the second field and the third field to a second electrical signal, and powering the implant using the second electrical signal. The second implant transducer and the first implant transducer are aligned.

[0007] In general, in another aspect, embodiments relate to another method. The method includes inducing, using an external transceiver, field pulses and receiving, by a transducer of an implant, the field pulses. Each of the field pulses has an excitation phase and a corresponding ringdown phase. Further, the implant is disposed up to a predetermined distance away from the external transceiver. The method further includes converting, using the transducer, the field pulses to electrical signal pulses and inducing, using the transducer, backscattered field pulses based on the field pulses. Each of the backscattered field pulses has the excitation phase and the corresponding ring down phase. The method still further includes sensing and recording uplink data, determining encoded uplink data by encoding the uplink data to align with the corresponding ringdown phase, and inducing, using the transducer, modulated backscattered field pulses by affecting the backscattered field pulses within the corresponding ringdown phase based on the encoded uplink data. The corresponding ringdown phase of each of the modulated backscattered field pulses includes at least three bits of the encoded uplink data. The method also includes receiving, by the external transceiver, the modulated backscattered field pulses, converting the modulated backscattered field pulses to modulated backscattered electrical signal pulses that include the encoded uplink data, and determining the uplink data by decoding the encoded uplink data from the modulated backscattered electrical signal pulses.

[0008] Other aspects and advantages of the claimed subject matter will be apparent from the following description and the appended claims.

BRIEF DESCRIPTION OF DRAWINGS

[0009] Specific embodiments of the disclosed technology will now be described in detail with reference to the accompanying figures. Like elements in the various figures are denoted by like reference numerals for consistency. The sizes and relative positions of elements in the drawings are not necessarily drawn to scale. For example, the shapes of various elements and angles are not necessarily drawn to scale, and some of these elements may be arbitrarily enlarged and positioned to improve drawing legibility. Further, the particular shapes of the elements as drawn are not necessarily intended to convey any information regarding the actual shape of the particular elements and have been solely selected for ease of recognition in the drawing.

[0010] FIGs. 1A-1C illustrate alignment between an implant and external coil in accordance with one or more embodiments.

[0011] FIG. 2A illustrates an external transmitter coil array in accordance with one or more embodiments.

[0012] FIG. 2B illustrates a power and magnetic field strength associated with an external transmitter coil array in accordance with one or more embodiments.

[0013] FIG. 3 shows a system in accordance with one or more embodiments.

[0014] FIG. 4 illustrates a model in accordance with one or more embodiments.

[0015] FIGs. 5 and 6 show a circuit diagram in accordance with one or more embodiments.

[0016] FIG. 7A shows a flowchart describing the method in accordance with one or more embodiments.

[0017] FIG. 7B shows a system in accordance with one or more embodiments. [0018] FIG. 8 illustrates magnetic field pulses in accordance with one or more embodiments.

[0019] FIG. 9 shows a system in accordance with one or more embodiments.

[0020] FIG. 10 shows a schematic of an uplink transmitter in accordance with one or more embodiments.

[0021] FIG. 11 shows a schematic of a peak detector in accordance with one or more embodiments.

[0022] FIG. 12 shows a schematic of a zero-crossing detector in accordance with one or more embodiments.

[0023] FIG. 13 A shows a schematic illustrating a phase shifter in accordance with one or more embodiments.

[0024] FIG. 13B shows another schematic illustrating a phase shifter in accordance with one or more embodiments.

[0025] FIG. 13C displays corresponding signals associated with the phase shifter in accordance with one or more embodiments.

[0026] FIG. 14 shows a schematic of a comparator in accordance with one or more embodiments.

[0027] FIG. 15A shows a schematic of a switched-array capacitor in accordance with one or more embodiments.

[0028] FIG. 15B displays corresponding signals associated with the switched-capacitor array in accordance with one or more embodiments.

[0029] FIG. 16 shows a recording analog-front end in accordance with one or more embodiments.

[0030] FIG. 17 displays time-varying electrical signals in accordance with one or more embodiments.

[0031] FIG. 18 shows a schematic of an external transceiver in accordance with one or more embodiments. [0032] FIG. 19 shows a model of an ME transducer in accordance with one or more embodiments.

[0033] FIG. 20A shows a flowchart describing the method in accordance with one or more embodiments.

[0034] FIG. 20B shows a system in accordance with one or more embodiments.

[0035] FIG. 21A and FIG. 21B show plots in accordance with one or more embodiments.

DETAILED DESCRIPTION

[0036] In the following detailed description of embodiments of the disclosure, numerous specific details are set forth to provide a more thorough understanding of the disclosure. However, it will be apparent to one of ordinary skill in the art that the disclosure may be practiced without these specific details. In other instances, well-known features have not been described in detail to avoid unnecessarily complicating the description.

[0037] Throughout the application, ordinal numbers (e.g., first, second, third, etc. may be used as an adjective for an element (i.e., any noun in the application). The use of ordinal numbers is not to imply or create any particular ordering of the elements nor to limit any element to being only a single element unless expressly disclosed, such as using the terms “before,” “after,” “single,” and other such terminology. Rather, the use of ordinal numbers is to distinguish between the elements. By way of an example, a first element is distinct from a second element, and the first element may encompass more than one element and succeed (or precede) the second element in an ordering of elements.

[0038] In the following description, any component described regarding a figure, in various embodiments disclosed herein, may be equivalent to one or more like-named components described regarding any other figure. For brevity, descriptions of these components will not be repeated regarding each figure. Thus, each and every embodiment of the components of each figure is incorporated by reference and assumed to be optionally present within every other figure having one or more like-named components. Additionally, in accordance with various embodiments disclosed herein, any description of the components of a figure is to be interpreted as optional embodiments which may be implemented in addition to, in conjunction with, or in place of the embodiments described regarding a corresponding like-named component in any other figure.

[0039] It is to be understood that the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to “a transducer” includes reference to one or more of such transducers.

[0040] Terms such as “approximately,” “substantially,” etc., mean that the recited characteristic, parameter, or value need not be achieved exactly, but that deviations or variations, including for example, tolerances, measurement error, measurement accuracy limitations and other factors known to those of skill in the art, may occur in amounts that do not preclude the effect the characteristic was intended to provide.

[0041] It is to be understood that one or more of the steps shown in the flowcharts may be omitted, repeated, and/or performed in a different order than the order shown. Accordingly, the scope disclosed herein should not be considered limited to the specific arrangement of steps shown in the flowcharts.

[0042] Methods of powering an implant when misaligned and transmitting data from the implant to an external transceiver are disclosed. The methods may be used separately or collectively by a disclosed system. Hereinafter, the method of powering the implant may be referred to as the active echo (AE) technique. Further hereinafter, the method of transmitting data from the implant to the external transceiver may be referred to as the pulse-width modulation (PWM) technique. The disclosed system includes an implant and external transceiver. In some embodiments, the size of the implant may be on the order of cubic millimeters. In some embodiments, the implant may be battery-free. When in use, the implant may be disposed or implanted within a user. The external transceiver may be disposed external to the user and up to a predetermined distance away from the implant.

[0043] Embodiments of the methods and system are disclosed below. However, a person of ordinary skill in the art will appreciate that the embodiments described below are not the only embodiments of the methods and system. Hereinafter, any “transducer” may be an electromagnetic transducer or electroacoustic transducer. Types of electromagnetic transducers include, without limitation, coils (such as those used in inductive coupling applications), antenna (such as those used in radiofrequency (RF) applications), magnetoelectric (ME) transducers, and ME film. Types of electroacoustic transducers include piezoelectric crystals (such as those used in ultrasound applications) and piezoelectric material. Electromagnetic transducers convert a magnetic field to an electrical signal and/or vice versa. Electroacoustic transducers convert an acoustic field to an electrical signal and/or vice versa. Hereinafter, the magnetic field or acoustic field is generically denoted a “field.” Accordingly, a “field pulse” may be a magnetic field pulse or acoustic field pulse. A “backscattered field pulse” may be a backscattered magnetic field pulse or backscattered acoustic field pulse. A “modulated backscattered field pulse” may be a modulated backscattered magnetic field pulse or modulated backscattered acoustic field pulse. As such, the AE technique may rely on magnetic field steering, beamforming, beam steering, etc. The PWM technique may rely on one or more high- quality-factor transducers. In view of this, one or more aspects of the disclosed methods and system may be modified to use and/or include a different type of transducer that may receive, emit/induce, and/or backscatter a different type of field without departing from the scope of the disclosure.

[0044] To power the implant, an omnidirectional wireless power transfer (WPT) platform is relied on that uses magnetic field steering. To detect the weak coupling between the implant and external transceiver as well as adaptively control the external transmitter coil array in a closed loop, the AE technique is used. Advantageously, the AE technique simultaneously enhances the robustness, efficiency, and charging range of miniature implants.

[0045] Bioelectronic implants (hereinafter also “implants”) have played a vital role in helping millions of patients with cardiovascular and neurological disorders since the first pacemaker developed in the 1950s. With recent advances in bioengineering, neuroscience, and medical research, implants are expected to further transform healthcare and medicine. Decades of effort have been dedicated to miniaturizing implants to reduce infection risks, simplify surgical procedures, reduce invasive implantation, improve patient acceptance, enhance target specificity, increase the operational lifespan of the implant, and open up new applications. Today, batteries and wires still dominate the volume of commercially available bioimplants. Replacing the batteries and wires with WPT technologies promises a significant leap forward in miniaturizing the implants to the cubic millimeter scale (hereinafter “millimetric”) or smaller. [0046] Despite the immense potential of wireless battery-less implants, two critical engineering hurdles must be overcome: ensuring safe and reliable WPT and establishing efficient bidirectional telemetry.

[0047] Fully-integrated, millimetric, and battery -free implants have been demonstrated with various WPT modalities (e.g., inductive coupling, radiofrequency (RF), ultrasound, light, and magnetoelectric (ME) modalities). However, each modality presents tradeoffs such as balancing the size of the transducer of the implant, the maximum power delivery under safety limits, and the misalignment tolerance between the implant and external transceiver. State-of-the-art millimetric implants have also been demonstrated for various applications with an implantation depth of up to five to six centimeters (cm) in a user, such as in cortical devices, spinal cord devices, vagus nerve devices, peripheral nerve stimulators, and leadless pacemakers. For applications that require more than ten centimeters of power transfer distance, such as in ingestible devices, replacing batteries with WPT technologies is more challenging because of safety limits on the maximum transmitter power of an external transceiver.

[0048] However, the power transfer efficiency (PTE) of WPT is known to severely degrade by misalignment between the external transceiver that powers the implant and the implant. Such misalignment is often inevitable during chronic implantation, due to body movements, respiration, heart beating, and limited control of implant orientation intraoperatively.

[0049] Three popular WPT modalities include inductive coupling, ultrasound, and ME, each of which have their own angular/rotational sensitivity. Inductive coupling is the most well-studied wireless charging method. However, it faces two maj or challenges when used in miniature implants that include the cubical reduction of the PTE with the radius of the receiver (e.g., a coil or other transducer) and increased specific absorption rates (SAR) in tissues of the user when adopting a higher frequency for coil size reduction. Ultrasonic WPT circumvents the user absorption issue due to millimetric transducers being acoustically resonant at frequencies on the order of megahertz (MHz) to efficiently deliver higher power to miniature implants. However, ultrasonic WPT suffers from reflections through different mediums and exhibits high sensitivity to rotations and lateral mismatch with a focused ultrasound transmitter. The ME WPT transducer converts an alternating current (AC) magnetic field to electrical voltage via acoustic vibration, which enjoys higher PTE than inductive coupling with millimetric receivers while avoiding the penetration issues in ultrasound. More importantly, thanks to the magnetic flux concentration effects, ME naturally offers superior misalignment tolerance.

[0050] Adaptation to the actual position of the implant is a clear pathway to address the misalignment tolerance. Closed-loop power regulation has been applied to inductive and ME WPT systems, where the implant reports its received power through uplink data back to the external transmitter for adjusting output power. However, this approach cannot handle severe lateral and rotational offsets and bums significantly higher power of the external transceiver at larger misalignment.

[0051] To alleviate the PTE reduction and robustness issues in all WPT systems (inductive, ultrasound, and ME), an intuitive solution is to adapt the magnetic field or acoustic field generated by the external transmitter of the external transceiver to better align with the receiver of the implant.

[0052] FIG. 1 A illustrates ideal alignment of an implant 100 relative to an external coil 105a of an external transceiver 110 in accordance with one or more embodiments. Specifically, FIG. 1A illustrates ideal alignment of an implant 100 relative to the external coil 105a as shown by the dashed vertical line when the implant 100 includes a magnetoelectric (ME) transducer (not shown), a type of transducer, as being perpendicular to the external coil 105a, another type of transducer. However, non-ME transducers, such as those used in ultrasound (e.g., piezoelectric crystals) and inductive coupling, may be used as an alternative. In some embodiments, non-ME transducers may be ideally aligned by being parallel to the external coil 105a. The term I#, where # = 1, 2, etc., is the current associated with each external coil 105a. The term H# is a portion of the magnetic field associated with each external coil 105a, where H_SUm is the sum of magnetic fields. However, a person of ordinary skill in the art will appreciate that each magnetic field is typically illustrated using multiple curved lines that emit radially out of and around each external coil 105a. This is not done in FIGs. 1 A-1C for viewing ease.

[0053] FIG. IB illustrates an angle of misalignment 112 of the implant 100 relative to the external coil 105a. A single external coil 105a does not possess the degree of freedom needed to adapt to the position of the implant 100. However, using multiple external coils 105a does possess the necessary degrees of freedom. Though FIG. IB illustrates a misalignment, specifically, a rotational offset, the implant 100 may be, alternatively or additionally, be misaligned by a translational offset. Hereinafter, the term “misalignment” includes at least one of a rotational offset or translational offset.

[0054] FIG. 1C illustrates that two external coils 105a can be used to manipulate the direction and intensity of the induced magnetic fields of the external coils 105a by adjusting the phase and amplitude of the current of each external coil 105a. This process is known as “magnetic field steering.” Hereinafter, two or more external coils 105a are collectively referred to as an “external transmitter coil array” 115, a type of external transmitter transducer array.

[0055] FIG. 2A illustrates placement of two external coils 105a, b (TXi and TX2) of an external transmitter coil array 115 in accordance with one or more embodiments. For example, in these embodiments, each external coil 105a, b has an outer diameter of 4.2 centimeters, inner diameter of two centimeters, and has 10 turns. However, in other embodiments, each external coil 105a, b may have any outer diameter, inner diameter, and number of turns without departing from the scope of the disclosure. In these embodiments, each external coil 105a, b is disposed in a plane where the planes are parallel to one another. Further, in these embodiments, the external coils 105a, b overlap with one another. In other embodiments, the external coils 105a, b may not overlap with one another.

[0056] FIG. 2B shows that the magnetic field and power of the external transmitter coil array 115 with two external coils 105a, b disposed at various positions relative to one another in accordance with one or more embodiments.

[0057] FIG. 2B shows that magnetic field steering of the two external coils 105a, b can efficiently operate up to a distance of five centimeters from the receiver (e.g., ME transducer) of the implant 100 when the diameter of each external coil 105a, b is 4.2 centimeters. By adding more external coils, the degree of freedom for magnetic field steering is increased, which leads to higher accuracy and broader coverage of steering the magnetic fields towards the receiver of the implant 100. This concept has been demonstrated on inductive WPT systems with various techniques for controlling the external transmitter coil array 115. However, all existing designs are developed for applications with larger receiver coils on the implant 100 (i.e., tens of centimeters in diameter) strongly coupled to external coils 105a, b of similar sizes, such as is used for phone charging. These designs cannot be adopted for millimetric implants 100 since the coupling is too weak to sense on the external transmitter coil array 115. Blindly searching for the optimal setting is possible with brute force or gradient algorithms, like blind adaptive beamforming, but will incur excessive latency, energy loss, and interruptions to normal operation. It also scales badly as the number of external coils increases.

[0058] FIG. 2B also shows that the two external coils 105a, b have opposing phases as shown by the arrow pointing left 202 and the same phases as shown by the arrow pointing right 204.

[0059] The AE technique harnesses this information to enable an omnidirectional WPT platform for misaligned millimetric implants. By actively transmitting or inducing an AE tone (hereinafter “first magnetic field”) using a coil on the implant 100 as illustrated in FIG. 3, the alignment between the external transmitter coil array 115 (TXi), i=l, 2, and transducer (RX) of the implant 100 can be indirectly sensed and controlled using magnetic field steering in a real-time closed loop.

[0060] FIG. 3 further illustrates an AE system 130 in accordance with one or more embodiments. The AE system 130 includes subsystems and parts configured to perform the AE technique. The AE system 130 includes the implant 100 and external transceiver 110. In practice, the implant 100 is disposed or implanted within a user 140. The external transceiver 110 is disposed external to the user 140 and up to a predetermined distance 145 away from the implant 100. The alignment of the implant 100 relative to the external transceiver 110 may be unknown. Accordingly, in some embodiments, the implant 100 may be aligned relative to the external transceiver 110. In other embodiments, the implant 100 may be misaligned relative to the external transceiver 110. In some embodiments, the predetermined distance 145 may be up to ten centimeters, inclusive.

[0061] The implant 100 includes an implant coil 120, a type of transducer. In some embodiments, an outer diameter of the implant coil 120 may be one centimeter or less. The implant coil 120 is electrically coupled to a coil transmitter 150 of a system-on-chip 155 (SoC) of the implant 100. The implant coil 120 is configured to induce, as prescribed by a controller 160 of the system-on-chip 155 of the implant 100 and using the coil transmitter 150, a first magnetic field. The implant 100 also includes an ME transducer 125. In these embodiments, the ME transducer 125 and implant coil 120 are aligned perpendicular to one another as FIG. 3 illustrates.

[0062] The external transceiver 110 includes the external transmitter coil array 115 that includes two or more external coils 105a-c. FIG. 3 specifically illustrates that the external transmitter coil array 115 includes three external coils 105a-c. In some embodiments, an outer diameter of each of the two or more external coils 105a-c may be at least two centimeters, inclusive. Each of the external coils 105a-c of the external transmitter coil array 115 receives the first magnetic field (not shown) induced by the implant coil 120. Accordingly, the first magnetic field propagates the predetermined distance 145. Each of the external coils 105a-c converts the first magnetic field to a first electrical signal. Accordingly, the external transmitter coil array 115 is acting as an external receiver coil array during this time interval.

[0063] In some embodiments, the implant 100 may have a volume of 14.2 cubic millimeters (mm³). The ME transducer 125 may include a five by two square millimeter (mm²) ME film. The system-on-chip 155 may be a fully-integrated low-power system-on- chip. Each external coil 105a-c may be electrically coupled to an external transmitter driver 165 (e.g., an off-the-shelf gallium nitride (GaN) driver) configured to drive each external coil 105a-c. Each external coil 105a-c may also be electrically coupled to a channel (CH#), where # =1,2,3, etc., of an external receiver chip 170 that features -161 dBm/Hz input- referred noise for sensing coupling coefficients. In some embodiments, the external receiver chip 170 may be part of the external controller 175. In other embodiments, the external receiver chip 170 may be separate and electrically coupled to the external controller 175. In some embodiments, the external transmitter driver 165 may be part of the external controller 175. In other embodiments, the external transmitter driver 165 may be separate and electrically coupled to the external controller 175. The external transmitter coil array 115 may be a mutual-inductance-canceled external transmitter coil array that achieves less than -50 dB S21 between each pair of external coils 105a-c to reduce resonant frequency shifts and resonant current reduction, where S21 is the s-parameter that represents voltage or power gain from one external coil 105a to another external coil 105b. The AE system 130 achieves a 6.8x PTE improvement over a single external coil 105a, b, or c rotationally misaligned by 90° from the ideal alignment and has less than 2% efficiency loss due to the AE tracking error. [0064] As for inductive coupling, the received power by an ME transducer 125 is proportional to the magnetic field along the long axis of the rectangular-shaped thin-film composite of the ME transducer 125. To increase PTE with adaptive magnetic field steering, the straightforward approach is to find the orientation of the ME transducer 125 of the implant 100 in real time (z.e., localizing) and use that information to calculate the current for each external coil 105a-c of the external transmitter coil array 115 needed to power the misaligned implant 100 using electromagnetic theories. Localizing the implant 100 is feasible with triangulation methods. However, triangulation methods can only measure the position of the implant 100 in three translation degrees of freedom (DoF), assuming no rotations, and requires extra computation power and time. Employing two or three orthogonal sensors on the implant 100 can achieve six DoF (z.e., rotation and translation) but makes the implant 100 bulky and consume more power and area to synchronize all data. Additionally, localizing adds extra burden to system communication.

[0065] However, the full 6-DoF information may not be necessary for guiding the magnetic field steering and AE technique for adaptive control using the external transmitter coil array 115. The idea is to actively transmit or induce a first magnetic field from the implant coil 120 of the implant 100. In turn, the external transmitter coil array 115 can receive the first magnetic field, convert the first magnetic field to a first electrical signal, and use the amplitude and phase of the first electrical signal to determine the amplitude of the current that each external coil 105a-c needs to induce to adequately power the misaligned implant 100. In these embodiments, where low-frequency near-field sensing is used, phase may be a polarity. Using the converse ME effect to produce the first magnetic field was an initial exploration. However, its complex and nonreciprocal energy conversion process leads to a low misalignment sensitivity and signal strength. These properties necessitate a higher performance receiver chain and calibration method to control the magnetic fields induced by the external transmitter coil array 115 that significantly complicates system analysis and loop closure. This calibration is also required for each ME film. Advantageously, in this disclosure, the implant coil 120 is added on the implant 100 to transmit or induce the first magnetic field to avoid these issues. By leveraging the reciprocity of coil coupling, the current that needs to be supplied to each external coil 105a- c to direct a magnetic field towards the misaligned implant 100 can be obtained without calculating or estimating the orientations of the implant 100. [0066] This principle is illustrated by modeling the implant coil 120 and external transmitter coil array 115 (TXi) as illustrated in FIG. 4. Based on the Kirchhoff s law, one can obtain: where Ri and Xi represent the resistance and reactance of each resonance tank, Ml i is the mutual inductance between TX1 and TXi, and Mi L is the mutual inductance between TXi and ME transducer 125 (RX). The RX side can be derived as: Equation (2)

The PTE is defined by the ratio of total alternating current (AC) power consumption of the external transmitter coil array 115 and the maximum AC power received by the ME transducer 125 with an optimal loading of: Equation (3)

The goal is to reduce power loss and increase efficiency, which can be achieved by increasing (PRecv/PLoss) according to equation (3) as:

Equation (4)

Using equation (2) to represent IL gives:

Equation (5)

Here, only the second term depends on the state of the ME transducer 125. The Cauchy- Schwarz Inequality can be applied to the numerator as:

Substituting equation (6) into equation (5) gives: Equation (7) where the equality exists only when (Mi /(Ri )l/2) =(Ri )l/2 li. If it is assumed that all external coils 105a-c of the external transmitter coil array 115 are identical and, thus, Ri of all external coils 105a-c are the same, the equality condition becomes:

E > E > > In

Equation (8)

^ML1 ^ML2 ^MLn

This is the condition needed to achieve high WPT efficiency. Since Mi L equals MLi in an electromagnetic induction system, the mutual inductance is proportional to the coupling coefficient. So, equation (8) can be further simplified to: Equation (9)

Equation (9) reveals the disclosed control strategy for magnetic field steering. The AE system 130 achieves peak efficiency when the ratio of currents allocated to the external coils 105a-c matches the ratio of their coupling coefficients with the ME transducer 125. This condition also implies that the magnetic field perpendicular to the implant coil 120 reaches its maximum. In addition to the amplitudes of the current for each external coil 105a-c, the phases of the current for each external coil 105a-c also need to be controlled. Since the wavelength of the resonance frequency (340 kHz) of the ME film is approximately 882 meters, i.e., significantly larger than the distance between the external transmitter coil array 115 and ME transducer 125, there is negligible phase difference between the first magnetic field received by each external coil 105a-c. Consequently, only identical or opposite polarities in AE sensing can be observed. On the other hand, when multiple external coils are involved in magnetic field generation, the magnetic field induced by each external coil will see a negligible phase difference within the space of interest. Therefore, the phase of the current applied to each external coil 105a-c of the external transmitter coil array 115 need only be the same or opposing phase relative to any external coil 105a, b, or c to efficiently add up or cancel their contributions for magnetic field steering purposes. This reference external coil 105a, b, or c may have the strongest current relative to the other external coils. Applying any other phase shifts will only consume higher external transmitter coil array power for the same result. For these reasons, only sensing and controlling two phases are needed. The polarities can be sensed and directly used at a phase and opposing phase for the external transmitter coil array 115. By aligning the long axis of an ME film (i.e., its ideal field direction) with the normal direction of the implant coil 120, it is ensured that the ME transducer 125 receives the strongest possible magnetic field from the external transmitter coil array 115 along its long axis under a given total external transmitter coil array power budget.

[0067] The AE technique effectively addresses the weak coupling challenge. The mutual inductance between the external coils 105a-c is another issue. To better analyze this, the driver voltage and current of the external transmitter coil array 115 is analyzed, where:

Vi = RiIi + ' k i j^M _ikI_k, Equation (10) where Vi and Ri are the output voltage of ith driver and the impedance of the resonance tank, respectively, Mik represents the mutual inductance between i-th external coil 105a-c and kth external coil 105a-c. Since the I may vary between each external coil 105a-c, the contribution of the second term may differ for each external coil 105a-c, which leads to a current-dependent resonant frequency shift and reduces the effective AC current in the external transmitter coil array 115. In an inductive coupling system, the operating frequency can be adjusted to compensate for the degradation of resonant current. However, the frequency in ME WPT must match the innate resonance frequency of the ME film (340 kHz) to achieve high efficiency. Therefore, a principled approach to optimize the relative position of multiple external coils 105a-c to minimize their mutual inductance is taken.

[0068] To explain the principle with two overlapping external coils 105a-c as shown in FIG. 2 A, the mutual inductance is: Equation (11) where M is the mutual inductance between two external coils 105a-c. Matching ^overlap and nonoveriap values with opposite signs will cancel the total flux and minimize their mutual inductance. Since = Bda, the distance between the two external coils 105a-c can be adjusted to change the overlapping area and achieve mutual inductance cancellation. The specific distance can be determined experimentally using a simulation and physical measurements. The simulation may be performed using the High-Frequency Structure Simulator (HFSS) software from Ansys Inc. or equivalent. By applying the same criterion to three external coils, the field control ability can be extended to 6 DoF and ensure that the mutual inductance is canceled for all pairs of external coils, resulting in the external transmitter coil array 115 that includes three external coils (hereinafter also “three-coil design”). Simulation results indicate that the mutual inductance in this three-coil design is even smaller than parallel external coils with no overlapping. Moreover, the overlapping external coil placement avoids regions with overly small magnetic field strength despite coil current optimization, such as the gap regions among nonoverlapping external coils.

[0069] Returning to FIG. 3, each external coil 105a-c is electrically coupled to the external transmitter driver 165, which is configured to drive each external coil 105a-c. The external coils 105a-c meet the proposed floor plan to cancel the mutual inductance in between. This phase initially powers on the implant 100 and charges a storage capacitor (Cstore) of the implant 100. The implant 100 recovers its clock signal (CLK) from the carrier frequency and, in some embodiments, decodes the downlink data (e.g., implant instructions) using the controller 160. To overcome the trade-off between the downlink data rate and transducer quality factor, a time-domain notch-based modulation scheme is used. The ME-induced voltage (hereinafter also “second electrical signal”) is rectified to Vrect and then converted by the DC-DC, where DC is direct current. The AE system 130 selects the higher voltage between V_rect and Vstim to power the low-dropout regulators (LDOs) and the rest of the AE system 130 to ensure the working stability, even if there are fluctuations in V_rect due to misalignment. After establishing the downlink communication, the controller 160 of the system-on-chip 155 of the implant 100 checks the device ID, programs the implant 100, and executes tasks sequentially. An on-chip physical unclonable function (PUF) is used to produce the device ID with minimum area and power to eliminate the need for nonvolatile memories and programming. The AE system 130 may be able to perform at least two tasks: programmable stimulation and AE sensing.

[0070] When the AE sensing task is enabled, the external transmitter driver 165 turns off all power transistors and enters a high-impedance mode. The coil transmitter 150 of the system-on-chip 155 of the implant 100 is activated to send out a few cycles of a single tone (z.e., “first magnetic field”) through the implant coil 120. Thanks to the low turn-off leakage of the external transmitter driver 165, there is little-to-no DC drift on the external receiver side when the external transmitter drivers 165 are off. Therefore, the first magnetic field is sensed or received by the external transmitter coil array 115 with high fidelity and fed into the external receiver chip 170 to determine the amplitude and phase of the first electrical signal associated with each external coil 105a-c. The external controller 175 uses the amplitude and phase of the first electrical signal associated with each external coil 105a-c to determine an amplitude of a current (z.e., output current) to be applied to each external coil 105a-c. The amplitude of the current may be controlled as described below.

[0071] The transfer curve between the duty cycle and output current is stored as a lookup table (LUT) in the external controller 175 (e.g., microcontroller unit (MCU)) because it is not perfectly linear. When the external controller 175 receives the amplitude and phase of the first electrical signal from the external receiver chip 170, the external controller 175 adjusts the output current to be proportional to the AE amplitude, and the current phase aligns with AE output. To ensure safety, the total power budget is fixed. This means that 1^ + 1^ + I is constant, which allows the required current value for each channel CH# of the external receiver chip 170 to be determined. Using the lookup table, the external controller 175 then prescribes what amplitude and phase of the current should be applied to each external coil 105a-c to induce a magnetic field or, in some embodiments, magnetic field pulses to adequately power the misaligned implant 100.

[0072] In some embodiments, the external transducers and implant transducers may be the ultrasonic transducers. Thus, when the AE technique is performed using electroacoustic transducers, the external coils 105a-c, implant coil 120, and ME transducer 125 may be replaced with ultrasonic transducers in some embodiments. To perform the AE technique, the implant ultrasonic transducer may emit/transmit one or more acoustic field pulses. Each of the at least two external ultrasonic transducers of the external transceiver 110 may receive the one or more acoustic field pulses and covert the one or more acoustic field pulses to the first electrical signal. The external receiver chip 170 may determine the received amplitude and phase of each of at least two first electrical signals. The received phases of the first electrical signals may be a related delay. The external controller 175 may determine the transmitted amplitudes and phases of the electrical driving signals and prescribe the transmitted amplitudes and phases of the electrical driving signals to the at least two external ultrasonic transducers to generate at least two acoustic fields. The acoustic transducer of the implant 100 receives the at least two acoustic fields and converts, using the system-on-chip 155, the at least two acoustic fields to the second electrical signal. The second electrical signal is used to power the implant 100. In some embodiments, each of the at least two acoustic fields may be an acoustic wave.

[0073] The AE technique is performed periodically to track the misalignment of the implant 100. The frequency at which to repeat AE technique is denoted the AE activation frequency, which represents how fast the AE system 130 updates the current applied to the external transmitter coil array 115. The AE activation frequency is programmable and application-dependent. For example, in a cardiac pacemaker implant that uses a 20 Hertz (Hz) sampling rate, which is around ten times larger than a heart rate, it is sufficient to track heartbeat-caused implant movement. During each AE sensing task, the ring voltage- controlled oscillator (VCO) in the coil transmitter 150 of the implant 100 is activated eight cycles in advance to stabilize. Considering the 16 cycles of AE transmission and the ring up/down time of the ME transducer 125, each AE sensing task takes less than 100 microseconds (ps). The interruption to ME power transfer by an AE sensing task is thus negligible.

[0074] To demonstrate the proposed WPT platform and further enhance system integration and portability, two chips are developed: an implantable biostimulator system- on-chip 155 with a technical emphasis on a three-level low-power coil transmitter 150, and a three-channel external receiver chip 170, including a compact, low-noise analog frontend, single slope analog-to-digital converter (ADC), and fast polarity detector.

[0075] FIG. 5 shows the low-power circuit diagram of the coil transmitter 150 configured to drive the implant coil 120 of the implant 100. The cycle controller is synchronized with the external transceiver 110 through the task trigger signal ENtrans, which is handled by the controller 160 of the system-on-chip 155 of the implant 100. When ENtrans goes high, the controller 160 activates the ring VCO, counts the number of output cycles, and stops the VCO after a programmable number of cycles. H-bridge power amplifier (PA) is chosen for its high-power efficiency with no static power loss. Using an electrical driving signal, such as a voltage, with a square wave variation to drive the H-bridge PA is a straightforward choice. However, the square wave variation includes undesirable harmonics, while only the fundamental tone injected into the inductor and capacitor (LC) tank (z.e., resonant circuit), behaving as a bandpass filter, is useful for AE coupling sensing. To save power, a multilevel PA concept to better mimic a sinusoidal signal is adopted that aims to reduce the turn-on time of the coil transmitter 150 and its power without compromising the amplitude of the fundamental tone.

[0076] For simplicity and compactness, a three-level PA is implemented by controlling the duty cycle and adding a nonoverlapping phase to the differential outputs of the VCO. During this phase, the outputs are shorted to a pre-charged reference voltage to form a current loop. Theoretically, there is no power consumption in this phase.

[0077] To produce the symmetrical VCO outputs with the same duty cycle but at 180° phase shift, a differential ring oscillator with a self-biased current tail is designed. The tail design acts as a current source without requiring additional gate biasing and behaves the same no matter with branch is on to thereby ensure identical common-mode output voltage and duty cycle for the differential outputs. The VCO outputs are then sent to the duty cycle control (DCC) circuit, which adjusts the duty cycle by changing the input biasing of the AC-coupled inverter. A three-bit voltage digital-to-analog converter (DAC) allows programming this biasing voltage from 0.1 to 0.5 Volts (V).

[0078] Regarding the external transceiver 110, conventionally, a separate pickup coil is used to receive uplink data from the implant 100. However, this separate coil and implant coil 120 are coupled and difficult to align, which will degrade the accuracy of coupling coefficient sensing in the AE technique. Meanwhile, this two-coil topology doubles the number of coils and complicates implementation.

[0079] FIG. 6 shows the architecture of the external receiver chip 170 in accordance with one or more embodiments. In some embodiments, the external coils 105a-c of the external transmitter coil array 115 may be reused for powering the implant 100 at different frequencies and times using magnetic fields. The magnetic fields induced by the external transmitter coil array 115 for power transfer have at 340 kHz frequency, while the first magnetic field is designed to have about a 1 MHz frequency. Advantageously, this scheme makes it easier to scale up the number of external coils 105a-c among the external transmitter coil array 115 for wider coverage. The AE sensing task is triggered by the external transceiver 110 to ensure synchronization. When the implant 100 transitions to the AE sensing task, the external transmitter driver 165 enters the high-impedance mode so that the first magnetic field received by the external coils 105a-c can be converted to a first electrical signal and received by the external receiver chip 170 without attenuation. After low noise amplification (LNA) and bandpass filtering, the first electrical signal VRX is sent to the peak detector (PD) and digitized by a single slope ADC as shown in FIG. 6. The first electrical signal is further amplified to rail-to-rail and sent to a digital polarity detector also shown in FIG. 6.

[0080] In the low-noise analog-front end (AFE), an inverter-based LNA and programable gain amplifier (PGA) are designed for good noise and programmable gain control. In some embodiments, the external receiver chip 170 may be overdesigned for noise performance to leave the signal-to-noise ratio (SNR) margin in the worst case, which ensures the AE served range is larger than the ME power range. The low gain mode ensures that the output of the LNA and PGA will not saturate when the implant 100 is very close to the external transmitter coil array 115 and the external coils 105a-c capture or receive a strong first magnetic field from the implant coil 120. In this LNA topology, inserting configuration switches at the first Gm stage to control the total gain will either occupy a large area to reach a low insertion resistor or degrade the total transconductance (Gm). Therefore, only the transimpedance in the second stage of LNA is changed to keep the total Gm constant. This leads to a near-constant input referred noise under different gain configurations. The first stage of PGA uses the same structure with a relaxed noise requirement. The second stage replaces the operational transconductance amplifier (OTA) with a resistance to handle a large voltage swing and maintain linearity. Bandpass characteristics consist of two passive high-pass filters and an active resistor-capacitor (RC) low-pass filter to suppress the interference from the external transmitter coil array 115 and narrow the noise band. LNA and PGA are biased by duplicated inverter structures to track process, voltage, temperature (PVT) variations. One drawback of the inverter topology is poor power supply rejection (PSR), which compromises noise performance and isolation between adjacent channels of the external receiver chip 170. To mitigate this issue, a local LDO for LNA and PGA individually in each channel of the external receiver chip 170 is employed. This design also ensures stability of the AFE by breaking the positive feedback loop from the supply of the LNA to the output of the front-end.

[0081] The PD tracks the peak voltage of VR_X (z.e., first electrical signal) and stores it on a capacitor. The single-slope ADC (SS-ADC) performs auto-zero on its comparator. The sampler waits a few cycles for the AE signal (z.e., first electrical signal) to ramp up and stabilize before starting the SS-ADC. During data conversion, a reverse ramp signal 1 from VDD to ground (GND) is supplied to the left plate of the sampling cap (Vca), then the counter starts. When V_cb reaches the threshold voltage of the inverter, the counter stops and outputs its value. Being on the external side, this counter is clocked by a high-quality off-chip clock source. The inverter-based comparator includes a dummy switch to compensate for the charge injection induced.

[0082] To sense the polarities of VR_X in all channels of the external receiver chip 170, an XOR logic can be performed on two amplified rail-to-rail VR_X signals. Inputs with identical polarities produce zero, while inputs with opposite polarities produce one or unity. Instead of comparing all pairs, any channel as a reference CLK can be selected and the relative polarities of all other channels determined. This is feasible because only the relative polarities for multicoil control are needed. However, sometimes not all channels receive a strong enough AE signal (z.e., first electrical signal) to be reliably amplified into a digital signal. Thus, selecting the channel receiving the strong AE signal as the reference clock is the most reliable way for the phase detection. To quickly find the strongest channel, a reverse ramp for SS-ADC is used so that higher amplitude signals take less time to finish the conversion. This design contributes to reliable and fast phase detection by minimizing the length of the AE sequence and thus saves power.

[0083] The external receiver chip 170 receives and outputs a series of time-varying signals. In some embodiments, channel 1 receives the strongest signal. After the reset signal is released, the PD first samples the peak value of each channel. The right plate of the sampling cap is auto-zeroing simultaneously. Then, the ADC starts conversion and concludes first for channel 1. As a result, the reference selection circuit immediately uses channel 1 as the reference clock for phase detection. Then, the XOR output will be sampled as the phase of each channel relative to channel 1. After the amplitude conversions of all channels finish, all six measurements are sent to the external controller 175 to determine the amplitude and phase for each of the external coils 105a-c of the external transmitter coil array 115. It is worth noting that only the relative polarities among channels are needed, rather than the absolute values. Thus, even if two or three channels stop at almost the same time and the clock selection circuit makes a mistake due to noise and variations, the sensed relative phase among channels is sufficient to ensure functionality.

[0084] FIG. 7A and FIG. 7B describes a method and illustrates an AE system 700 that relies on the AE technique in accordance with one or more embodiments. By performing the following method, the external transceiver 705 may wirelessly power the implant 710 no matter the angle of misalignment 715 of the implant 710 relative to the external transceiver 705. It is to be understood that one or more of the blocks shown in the flowchart may be omitted, repeated, and/or performed in a different order than the order shown. Further, one or more of the blocks may be performed substantially in series or in parallel (z.e., simultaneously). Accordingly, the scope of the disclosure should not be considered limited to the specific arrangement of blocks shown in the flowchart.

[0085] In practice, the implant 710 may be disposed or implanted within a user (not shown in FIG. 7B) prior to performing the method described in FIG. 7A. The external transceiver 705 may be disposed up to a predetermined distance 720 away from the implant 710 and external to the user. The misalignment of the implant 710 relative to the external transceiver 705 may be unknown and/or change over time. As such, the implant 710 may be aligned relative to the external transceiver 705 for one or more time intervals.

[0086] In block 725, a first field 730 is induced using or emitted from the first implant transducer 735 of the implant 710. To do so, the controller of the implant 710 (not shown in FIG. 7B) may prescribe an amplitude and phase of an electrical signal to flow to or through the first implant transducer 735 such that the first implant transducer 735 emits the first field 730. In some embodiments, the coil transmitter may be designed as illustrated in FIG. 5. It will be understood that as used herein the terms “electrical signal” and “electrical driving signal” refer to a voltage or current.

[0087] In block 740, the first field 730 is received by the external transmitter transducer array 745 of the external transceiver 705 as shown by arrows 750a, b. The external transmitter transducer array 745 includes at least two external transducers 752a, b. As such, each of the at least two external transducers 752a, b receives the first field 730. Accordingly, the first field 730 propagates the predetermined distance 720 between the external transceiver 705 and the implant 710.

[0088] In block 755, the first field 730 is converted to a first electrical signal for each of the at least two external transducers 752a, b. Accordingly, if the external transmitter transducer array 745 includes two external transducers 752a, b, then there are two first electrical signals. [0089] In block 760, a received amplitude of the first electrical signal and a received phase of the first electrical signal is determined for each of the at least two external transducers 752a, b. In some embodiments, an external receiver chip (not shown in FIG. 7B) of the external transceiver 705 determines the received amplitude and received phase of the first electrical signal associated with each of the at least two external transducers 752a, b. In some embodiments, the external receiver chip may be designed as illustrated in FIG. 6.

[0090] In block 765, a first transmitted amplitude of a first electrical driving signal and a first transmitted phase of the first electrical driving signal is determined using the received amplitude and the received phase of the first electrical signal independently for each of the at least two external transducers 752a, b. A second transmitted amplitude of a second electrical driving signal and a second transmitted phase of the second electrical driving signal is determined using the received amplitude and the received phase of the first electrical signal independently for each of the at least two external transducers 752a, b. In some embodiments, an external controller 770 of the external transceiver 705 may perform block 765.

[0091] In block 775, a first external transducer 752a among the at least two external transducers 752a, b induces a second field 780 by applying the first transmitted amplitude of the first electrical driving signal at the first transmitted phase to the first external transducer 752a. A second external transducer 752b among the at least two external transducers 752a, b induces a third field 785 by applying the second transmitted amplitude of the second electrical driving signal at the second transmitted phase to the second external transducer 752b. The second field 780 and third field 785 are induced substantially simultaneously. In some embodiments, the second field 780 and third field 785 may be field pulses.

[0092] In block 790, a second implant transducer 795 of the implant 710 receives the second field 780 and the third field 785 as shown by arrow 750c. Accordingly, the second field 780 and the third field 785 propagate the predetermined distance 720 between the external transceiver 705 and the implant 710.

[0093] In block 796, the second implant transducer 795 converts the second field 780 and the third field 785 to a second electrical signal. [0094] In block 797, the implant 710 is powered using the second electrical signal. In some embodiments, a power manager (not shown in FIG. 7B) of the implant 710 may power the implant 710. In embodiments where the second field 780 and third field 785 are field pulses, the power manager may intermittently power the implant 710 using the second electrical signal.

[0095] Turning to the PWM technique, the PWM technique may be used to transmit data from the implant 100 to the external transceiver 110. For any transducer, there exists an intrinsic tradeoff between PTE and communication bandwidth. While a high-quality- factor transducer is ideal for maximizing PTE, it inherently reduces communication bandwidth. The PWM technique addresses this limitation by encoding multiple bits of uplink data within a single corresponding ringdown phase as described below. The PWM technique may encode uplink data sensed and recorded by a sensor of the implant 100 within backscattered magnetic field pulses emitted from the transducer. Advantageously, the PWM technique may transmit at least three bits of encoded uplink data within each corresponding ringdown phase of each backscattered magnetic field pulse. The PWM technique may be performed, in part, using an energy excitation (EE) circuit disposed within the implant 100. The EE circuit may include a switched-inductor or switched- capacitor array configured to perform an EE technique. The EE circuit rapidly extracts and dissipates kinetic energy within the transducer of the implant 100 during the corresponding ringdown phase of each backscattered magnetic field pulse emitted by the transducer in response to receiving magnetic field pulses from the external transceiver 110. In doing so, the uplink data may be encoded within the backscattered magnetic field pulses during the corresponding ringdown phase within the time-domain. Advantageously, the disclosed PWM technique improves the reliability and data rate of uplink telemetry with an ME transducer. The ME transducer is a composite laminate made with coupled magnetostrictive and piezoelectric films. The piezoelectric layer converts acoustic fields to electrical fields. However, any high-quality-factor transducer may be used to perform the PWM technique as high-quality-factor transducers have relatively long corresponding ringdown phases. It will be understood that while illustrated for backscattered magnetic field pulses, other backscattered field pulses are contemplated such as when the fields are acoustic fields. [0096] An emerging WPT technology using magnetoelectric (ME) effects stands out due to its superior PTE using a transducer on the order of cubic millimeters, high-power delivery within safety limits, and lower sensitivity to misalignment. The ME effect converts low-frequency magnetic fields to electrical signals via acoustic coupling. This low-operating frequency results in reduced body absorption and reflection to allow deeper tissue penetration compared to other WPT methods. Previous work with ME power transfer shows that an operating depth up to 60 mm (0.1 mT magnetic field) can be achieved without violating the electrical field limit and the specific absorption rate (SAR) limit in IEEE standard. Furthermore, as attributed to the high permeability of the magnetostrictive materials, the magnetic flux concentration effect greatly enhances robustness to angular/rotational misalignment. These features make ME a compelling WPT solution for millimeter-scale implants 100.

[0097] Bidirectional wireless telemetry also is essential for real-time sensing and closed-loop physiological control of the implant 100. Further, bidirectional wireless telemetry must be integrated with the WPT scheme to achieve optimal performance and device miniaturization. Integrating a second antenna, most commonly an inductive coil on the implant 100, is the most straightforward uplink solution with a high data rate. However, this approach increases the size of the implant 100, complicates the assembly of the implant 100, and introduces interference between power and telemetry channels. Instead, sharing a single transducer for power and data transfer is highly desirable for device miniaturization and system efficiency. For common WPT modalities, such as inductive coupling and ultrasound, low-power backscatter communication techniques can be readily applied by modulating the load impedance to the transducer of the implant 100. Uplink communication with ME transducers is more challenging due to its non-reciprocal properties. Recent studies reveal possible solutions based on converse ME effects, which refers to the phenomenon where vibrations in the magnetostrictive layer lead to changes in the material magnetization. In one study, uplink telemetry is realized by actively driving an ME transducer. Specifically, the ME transducer is driven with an electrical driving signal having a time-varying value made up of two cycles of sine wave variations with opposite phases and pulse position encoding is used for data modulation. In another study, the active driving scheme is extended to high-frequency resonance modes to demonstrate simultaneous power transfer and uplink transmission. While active driving achieves a high data rate, driving the kQ-impedance of ME transducers with large amplitude consumes significant power to generate a strong enough received signal. In still another study, researchers discovered that modulating the capacitive or resistive loading of an ME transducer during its corresponding ringdown phase modulates its vibration frequency without burning power to effectively achieve ultra-low-power backscatter communication. Based on this principle, an ME backscatter telemetry using load shift keying (LSK) is demonstrated. In experiments, the high-Q (quality factor) ME transducer continues to vibrate for more than 30 cycles after turning off the excitation magnetic field pulses, during which the capacitive loading is switched according to uplink data. Despite the advancement, LSK ME backscatter encounters limited SNR and data rate. Due to the high- Q nature of the ME transducer, a direct trade-off between frequency shift and signal strength caps the SNR achievable at the receiver end. Besides, the excitation and ringdown processes take tens of ME cycles to transmit a single bit, which limits the achievable data rate.

[0098] FIG. 8 illustrates magnetic field pulses 800 used in the disclosed PWM technique in accordance with one or more embodiments. As previously described, the AE system 130 includes the implant 100 and external transceiver 110. The implant 100 is implanted or disposed within the user 140 prior to performing the PWM technique. The external transceiver 110 is disposed external to the user 140 and up to a predetermined distance 145 away from the implant 100. The external transceiver 110 includes an external transmitter coil 805 and external receiver coil 810. If the AE system 130 is also configured to perform the AE technique as previously described, the external transmitter coil 805 may be replaced with the external transmitter coil array 115.

[0099] The external transmitter coil 805 is configured to induce or emit the magnetic field pulses 800. Each of the magnetic field pulses 800 has an excitation phase 815 and corresponding ringdown phase 820. An ME transducer (not shown in FIG. 8) of the implant 100 receives the magnetic field pulses 800 and converts the magnetic field pulses 800 into electrical signal pulses (not shown in FIG. 8). In response to the ME transducer receiving the magnetic field pulses 800, the ME transducer may induce backscattered magnetic field pulses 825. Each of the backscattered magnetic field pulses 825 has the excitation phase 815 and corresponding ringdown phase 820. In some embodiments, downlink data may be transmitted from the external transceiver 110 to the implant 100 by being encoded within the magnetic field pulses 800. In the PWM technique, uplink data is transmitted from the implant 100 to the external transceiver 110 by being encoded within one or more backscattered magnetic field pulses 825 within the corresponding ringdown phase 820.

[0100] Achieving robust and high-data-rate PWM within the backscattered magnetic field pulses 825 involves two key aspects. One, the EE circuit electrically coupled to the ME transducer is needed to rapidly dissipate stored energy in the ME transducer. This process leads to a significant and quick amplitude reduction of the backscattered magnetic field pulses 825 to indicate the change in pulse width. As a result, the “on” and “off’ states of the backscattered magnetic field pulses 825 are clearly distinguishable to thereby enhance the reliability and accuracy of data transmission with PWM. Two, precise timing control is implemented to accurately reduce the amplitude of the backscattered magnetic field pulses 825 to enable PWM with a duty cycle that is modulated by the encoded uplink data to be transmitted from the implant 100 to the external transceiver 110. This allows for the effective encoding of multiple bits of uplink data into the corresponding ringdown phase 820 of each backscattered magnetic field pulses 825, which enhances the effective data rate for neural recording purposes.

[0101] FIG. 9 illustrates a PWM system 900 in accordance with one or more embodiments. The PWM system 900 includes subsystems and parts configured to perform the PWM technique. In some embodiments, the PWM system 900 may include the AE system 130 described relative to FIG. 3. The PWM system 900 includes the implant 100 and external transceiver 110.

[0102] The external transceiver 110 includes an external transmitter 165a, external controller 175, and external receiver 170a. The external transmitter 165a may include the external transmitter coil 805. The external receiver 170a may include the external receiver coil 810. In some embodiments, as illustrated in FIG. 9, the external receiver coil 810 may be disposed within the external transmitter coil 805.

[0103] The implant 100 includes the ME transducer 125 and system-on-chip 155. In some embodiments, the implant 100 also includes one or more electrodes, such as a stimulating electrode 830 and recording electrode 835. As such, the system-on-chip 155 may include a stimulator 840 and neural recording module 845. In these embodiments, the PWM system 900 may be used as a wireless neural recording system. In these embodiments, the stimulating electrode 830 is configured to stimulate the user 140. The recording electrode 835 is configured to sense user data 850 caused in response to the stimulation. The neural recording module 845 is configured to record the user data 850 and, in some embodiments, local field potentials within a brain of a user 140. Specifically, local field potentials are a measure of the electrical activity of a group of neurons in the brain, which may be measured by one or more electrodes. Accordingly, the user data 850 is associated with the user 140. In some embodiments, the system-on-chip 155 includes a sensor 855. The sensor 855 may be configured to sense and record sensing data 860, such as temperature data and/or voltage data. The sensing data 860 may be associated with the user 140 and/or implant 100. As such, the system-on-chip 155 may include a temperature sensing circuit and/or voltage sensing circuit (neither of which is shown in FIG. 9). The user data 850, temperature data, and/or voltage data (collectively sensing data 860) are generically denoted “uplink data” 865 hereinafter. In some embodiments, the uplink data 865 may also include data associated with a state of the implant 100, such as power consumption of the implant 100 over time.

[0104] For wireless reception of power and the programming of data, the system-on- chip 155 of the implant 100 includes the power manager 180 and downlink data recovery module (not shown) electrically coupled to the ME transducer 125. As such, the magnetic field pulses 800 emitted from the external transceiver 110 are used to power the implant 100 and, in some embodiments, transmit downlink data 866 as shown by the arrow 867.

[0105] To transmit the uplink data 865 from the implant 100 to the external transceiver 110 in the direction shown by arrow 868, the system-on-chip 155 includes an uplink transmitter 870 with EE circuit 875, PD 880, and delay control 885. The uplink transmitter 870 is configured to encode the uplink data 865 such that the ME transducer 125 can transmit the encoded uplink data to the external transceiver 110 via the backscattered magnetic field pulses 825. To do so, the modulation scheme of the PWM technique is characterized by substantial amplitude reductions in the backscattered magnetic field pulses 825 at different time points within the corresponding ringdown phase 820. Following the modulation scheme, modulated backscattered magnetic field pulses are received by the external receiver coil 810. The external receiver coil 810 converts the modulated backscattered magnetic field pulses to modulated backscattered electrical signal pulses. The backscattered electrical signal pulses are sent to the external receiver 170a, which incorporates an AFE (not shown in FIG. 9) for amplification and filtering. The modulated backscattered electrical signal pulses are processed using algorithms that include envelope extraction and drop detection functions for data demodulation.

[0106] In some embodiments, demodulation involves peak voltage drop detection within the backscattered magnetic field pulses. This approach detects voltage drops in the electrical signal pulse 890 at different time points to recover the pulse-width. This demodulation method works well with limited computing resources, which makes it suitable for implementation in commercial MCUs. In other embodiments, demodulation involves a multilayer perceptron (MLP). While more complex than peak voltage drop detection, the MLP offers enhanced robustness and a lower bit error rate.

[0107] In some embodiments, the PWM system 900 includes bio-stimulation, neural recording, temperature sensing, and bidirectional communication as mentioned above. To power the PWM system 900, the electrical signal pulses 890 are rectified and regulated by the power manager 180 to generate a stable 1 V supply voltage. A DCDC converter is used to generate the stimulation voltage VSTIM used by the stimulator 840. The supply of the LDO is connected to the higher voltage between VRECT and VSTIM to guarantee enough voltage headroom. To ensure robust downlink and uplink operation, a decoupling capacitor of 500 nanofarads (nF) may be placed at the output of the rectifier to ensure a stable VRECT during the corresponding ringdown phase 820. In some embodiments, the encoded downlink data (encoded in the magnetic field pulses 800 and, thus, electrical signal pulses 890) is received by the downlink receiver 902, decoded to generate the downlink data 866, and sent to the controller 160. The downlink data 866 may include stimulator parameters and other SoC configurations needed for programming.

[0108] When the PWM system 900 is in uplink mode, the uplink transmitter 870 may encode the uplink data 865 and transmit the encoded uplink data 871 to the ME transducer 125. This transmission is enabled by the EE circuit 875 and PD 880 of the uplink transmitter 870, which allows for rapid amplitude reduction of the electrical signal pulses 890 at specific moments. An asynchronous first in, first out (FIFO) serves as a buffer for handling any mismatch between uplink data rates, sensing data rates, and user data rates. During uplink mode, the external transmitter 165 a of the external transceiver 110 is turned off frequently to create the corresponding ringdown phase 820, which interrupts power transfer and roughly halves the received power. However, thanks to the low-power consumption of the uplink telemetry and system-on-chip 155, the delivered power is enough to sustain the operation of the PWM system 900.

[0109] The EE circuit 875 is configured for an EE technique used within the PWM technique. In some embodiments, the EE circuit 875 may include a switch ed-inductor and switched-capacitor array. To enable robust PWM with a substantial amplitude change that accurately represents the presence or absence of the backscattered magnetic field pulses 825, it is necessary to dissipate the stored energy in the ME transducer 125. Additionally, to minimize the time of the corresponding ringdown phase 820 for improved uplink data rate transmission, it is desirable to dissipate the stored energy quickly. By applying the EE technique to the electrical signal pulses 890 during the corresponding ringdown phase 820, the stored energy in the piezoelectric material of the ME transducer 125 can be rapidly extracted for dissipation purposes, which results in a significant reduction in the amplitude of the backscattered magnetic field pulses 825.

[0110] The EE circuit 875 is electrically coupled to the ME transducer 125 when the output voltage of the ME transducer 125 reaches its peak, which corresponds to the maximum stored energy in the piezoelectric capacitor Cp of the ME transducer 125. Upon connection, the energy stored in the piezoelectric capacitor Cp is transferred to the EE circuit 875 to cause a flip in the voltage of the ME transducer 125. Once the flip is complete, the EE circuit 875 is disconnected from the ME transducer 125 and the energy in the EE circuit 875 is promptly removed by disconnecting or shorting the terminals of the piezoelectric capacitor Cp.

[oni] The PWM technique induces a detectable amplitude reduction in the output voltage of the ME transducer 125 (z.e., electrical signal pulses 890) at an exact time during the corresponding ringdown phase 820. To achieve a high SNR, the amplitude reduction must be sharp and the residual voltage minimal. Since the output voltage of the ME transducer 125 is induced by the acoustic vibration of the piezoelectric layer of the ME transducer 125, reducing the amplitude necessitates extracting and dissipating the mechanical kinetic energy stored in the ME transducer 125. Simply shorting the two terminals of the ME transducer 125 is ineffective, as it will cause the ME transducer 125 to enter short circuit resonance, which leaves a significant amount of energy in the resonator that continuously emits signals. This target aligns with the goal of piezoelectric energy harvesting (PEH) circuits, which maximizes the extraction of kinetic energy from a piezoelectric transducer. Therefore, inspiration is drawn from state-of-the-art PEH interface designs to develop the EE circuit 875 for uplink telemetry.

[0112] Conventional PEH circuits, such as synchronized switch harvesting on inductor (SSHI) and synchronous electrical charge extraction (SECE) designs, require bulky off- chip inductors to achieve good energy extraction performance. Moreover, zero-current detection in the EE technique will consume significantly more power with ME transducers 125 that operate at much higher frequencies (hundreds of kHz) than typical PEH interfaces (<kHz). In contrast, the synchronized switch harvesting on capacitors (SSHC) technique achieves comparable performance by employing a series of capacitors to flip the voltage across the ME transducer 125 synchronously. Based on the principle of SSHC, the EE circuit should have comparable capacitance as the output capacitor of the ME transducer 125, which can be integrated on-chip (~1 nF). Since SSHC is based on charge sharing, the requirements for on-resistance of the switches and timing precision are less stringent compared to that of the inductor counterpart, making it more suitable for implants 100 with ME transducers 125.

[0113] In some embodiments, ultrasonic transducers are used in the external transceiver and in the implant. Thus, in some embodiments, when the PWM technique is performed using electroacoustic transducers, the external transmitter coil 805, external receiver coil 810, and ME transducer 125 may be replaced with ultrasonic transducers. To transmit the uplink data 865 from the implant 100 to the external transceiver 110, the uplink transmitter 870 with EE circuit 875, PD 880, and delay control 885 transmit the encoded uplink data 871 via backscattered acoustic field pulses. The modulation scheme of the PWM technique is characterized by amplitude reductions in the backscattered acoustic field pulses at different time points within the corresponding ringdown phase 820. The modulated backscattered acoustic field pulses are received by the external transmitter ultrasonic transducer. The external transmitter ultrasonic transducer converts the modulated backscattered acoustic field pulses to modulated backscattered electrical signal pulses. The modulated backscattered electrical signal pulses are sent to the external receiver 170a. The modulated backscattered electrical signal pulses are processed using algorithms that include envelope extraction and drop detection functions for data demodulation. [0114] FIG. 10 illustrates the uplink transmitter 870 of the system-on-chip 155 of the implant 100 in accordance with one or more embodiments. The circuits of the uplink transmitter 870 include the EE circuit 875, PD 880, and delay control 885. The circuits work synergistically to enable efficient and high-data-rate PWM for wireless uplink data transmission. The pulse duration control mechanism relies on an ME cycle counter 905 that monitors the number of ME cycles during the corresponding ringdown phase 820. ME cycles are defined with an ME clock recovery circuit. This circuit compares the two voltages of transducer terminals. When the two voltages cross, the circuit is triggered, indicating the start or end of the ME cycle. The ME cycle counter 905 reduces the amplitude of the electrical signal pulses 890 only when the counting number matches the predefined reference, which is encoded in the uplink data 865 intended for transmission. This meticulous timing control allows for encoding at least three bits within a single corresponding ringdown phase 820 of the backscattered magnetic field pulses 825. Accordingly, the PWM technique advantageously has improved uplink data rates compared to other uplink data transmission methods. To optimize the amplitude change (z.e., modulation) of the backscattered magnetic field pulses 825 and achieve a high SNR, the EE technique is performed precisely at the peak voltage of the ME transducer 125. This approach maximizes the distinction between the “on” and “off’ states of the backscattered magnetic field pulses 825 to thereby enhance data transmission reliability.

[0115] FIG. 11 shows the block diagram of the PD 880 and related modules in accordance with one or more embodiments. When a corresponding ringdown phase 820 is detected during the uplink phase, the PD 880 will determine when to modulate the voltage of the electrical signal pulses 890. The ENZCD and ENDPS electrical signals are sent to the PD 880 to enable the zero-crossing detector (ZCD) 910 and phase shifter 915. When the ZCD 910 detects the zero-crossing point Pzc, the PS 915 will shift the electrical signal pulse 890 by TME/4 and 3TME/4 to find the peak points to perform the voltage flipping using the folded delay chain 920.

[0116] To optimize the efficacy of EE circuit 875, it is crucial to establish a connection between the EE circuit 875 and ME transducer 125 precisely when the energy stored in the piezoelectric capacitor Cp of the ME transducer 125 reaches its peak value. This necessitates the design of a high-accuracy PD 880 that is also power efficient to align with the ultra-low power nature of magnetic field pulses 800. Detecting the zero crossing of the output voltage of the electrical signal pulse 890 of the ME transducer 125 is a simpler task compared to detecting the peak value of the output voltage. This can be achieved using a voltage comparator (hereinafter “comparator”) with its inputs directly connected to the two terminals of the ME transducer 125. This eliminates the need for an external reference. Furthermore, zero-voltage detection offers improved robustness against noise and offset in the comparator. This is because at the zero crossing, the output voltage exhibits the steepest slope and the largest change within a given time. In the case of a sinusoidal electrical signal pulse 890, there exists a 90° phase shift between its peak value and zero value. Taking advantage of these observations, the ZCD 910 and PS 915 are combined to detect the peak value of the output voltage of the ME transducer 125.

[0117] The PD 880 has two phases: a calibration phase 930 and data transmission phase 935. At the beginning of each uplink session, the calibration phase is performed to convert the period of the electrical signal pulses 890 TME to digital codes. The calibration phase is performed during the corresponding ringdown phase 820 to avoid inaccurate timing caused by the mismatch between the driving frequency of the uplink transmitter 870 and the selfresonance frequency of the ME film of the ME transducer 125. After calibration, the uplink transmitter 870 enters the data transmission phase and the PD 880 produces peak-aligned pulse trains (ENSCEE) for controlling the EE. The phase shifter of the PD 880 adds TME/4 and 3TME/4 delays to the output electrical signal pulse of the zero-crossing detector (ZCD) to locate positive and negative peaks. The delay control 885 tracks the number of corresponding ringdown phases 820 and ME cycle counter 905 over the entire uplink session. The delay control 885 also activates the PD 880 for four ME cycles at specific moments based on uplink data 865. Within these four cycles, the ZCD is activated only in the first cycle while the phase shifter generates eight ENSCEE pulses based on its calibrated delay line to lower the average power to 15 nanowatts (nW).

[0118] In the PD 880, TME/4 and 3TME/4 delays are required to derive the peak points from the zero-crossing point. Conventionally, shift register and multiplier are necessary to calculate the result, which is area and power-consuming, especially considering the 12-bit TDC output. In this disclosure, the PS 915 employs a shift-only arithmetic to approximate the TME/4 and 3TME/4 delays with minimum overheads.

[0119] FIG. 12 illustrates a schematic of the ZCD 910 in accordance with one or more embodiments. The ZCD 910 is enabled at a specific cycle during every corresponding ringdown phase 820 in uplink mode according to the uplink data 865. The accuracy of the ZCD 910 directly affects the efficacy of the EE circuit 875. For accurate detection, the ZCD 910 uses a comparator with an intentionally added offset current to reduce detection delay by intentionally adding an offset without significantly increasing the biasing current. However, using a fixed sinking current Ii leads to overcompensation during later ME ringdown cycles when the signal amplitude is smaller. To address this issue, an adaptive sourcing current E that increases with each cycle to complement Ii is introduced. Simulation results demonstrate that incorporating the adaptive current greatly reduces the ZCD delay and ensures more consistent performance throughout the entire corresponding ringdown phase 820. E is biased by an ME envelope extractor 925 that is also necessary for downlink, which eliminates the need for additional hardware.

[0120] FIG. 13 A and FIG. 13B illustrate a circuit diagram of a PS 915 in accordance with one or more embodiments. The PS 915 may be specifically referred to as a digital phase shifter (DPS). The PS 915 incorporates time-to-digital calibration and delay-lockingbased pulse generation for peak detection. For a sinusoidal signal with a constant frequency, the positive and negative peaks exhibit phase shifts of 90° and 270°, respectively. These phase shifts correspond to a time delay of one-quarter and three- quarters of a cycle in the time domain. In the case of the backscattered magnetic field pulses 825, it maintains a constant frequency, which is its mechanical resonance frequency. While the EE technique for amplitude modulation can alter the voltage of the ME transducer 125 to thereby shift the position of the zero-crossing points, it does not affect the positions of the peaks. As such, a calibration phase and a delay -locking phase are used to attain phase shifting of zero-crossing for precise peak detection. The calibration process is autonomously conducted during the initial corresponding ringdown phase 820 to determine the period of the backscattered magnetic field pulses 825. This period acts as a reference for generating the desired delays in the subsequent delay-locking phase. Since the resonance frequency of the backscattered magnetic field pulses 825 remains unchanged, a one-time calibration is sufficient to transmit all of the encoded uplink data 871 from the implant 100 to the external transceiver 110. Two zero-crossing detectors are employed to define an enable signal, ENCALL The core function of the calibration phase involves a time- to-digital conversion process. Traditional time-to-digital converters (TDC) often face a tradeoff between resolution and size that requires many delay stages to achieve higher resolution. To overcome this challenge, a folded delay chain that includes a ring oscillator (RO) structure is used. This design enables the recycling of delay stages for a high resolution of 8 nanoseconds (ns) and improved efficiency in terms of area and energy consumption. The oscillation of the RO is triggered by a start pulse synchronized with the rising edge of ENCALL The delay counting is carried out cooperatively by a cycle counter and a stop stage monitor. The cycle counter monitors the RO output YB<3> and provides a 5-bit output CNT<4:0>, which indicates the number of oscillation cycles. The stage monitor senses the delay propagation within one cycle. At the falling edge of ENCALI, the values of CNT<4:0> and STG<4:0> are stored in a register, which serves as a reference for locking the delay and generating the pulses that indicate peak detection. Advantageously, this self-calibration scheme demonstrates robustness against process variations, supply voltage fluctuations, and temperature changes to ensure reliable performance in varying operating conditions.

[0121] The PS 915 comprises a differential 4-stage folded delay chain with a 15 ns step (z.e., 0.5% of TME), a cycle counter, sampling D Flip-Flops (DFFs), and a digital pulse generator. The folded delay chain is used to mitigate the exponentially-increasing area requirements for higher time resolutions. During the calibration phase, the PS 915 works as a TDC to record the time interval between two zero-crossing points. The output of the TDC comprises two parts: the output of the cycle counter (DCYC<4:0>, binary) and the output of the stage counter (DSTG<6:0>, thermometer). The digital code is stored in the sampling DFFs. At the data transmission phase, the PS 915 is triggered when the output of the cycle counter matches the desired delay such that pulsing signals will be sent to the EE interface to execute the modulation.

[0122] Shift-only arithmetic is employed to derive the TME/4 and 3TME/4 delay from the TDC output code of the TME, which saves considerable computation without sacrificing much accuracy. The digital code of TME can be represented by:

TME = delay X [8 X dec{D_CYC < 4: 0 >} + dec{D_STG < 6: 0 >}], Equation (12) where dec{D} represents the decimal format of the digital code D. When divided by 4, the first half of equation (12) can be shifted right by 2 bits, resulting in 8 X dec(D_CYC < 4: 2 >} plus 2 x {D_CYC < 1: 0 >}. The second half, which consists of the thermometer part from the stage counter, can be approximated by D_STG < 3 > when divided by 4. This means that when D_STG < 6: 0 > equals or is larger than 0001111 (thermometer), the division result is set to 1; when it is smaller, the result is set to 0. Therefore, the final TME/4 can be expressed as: 3 >}, Equation (13) where {D_CYC < 1 >, D_CYC < 0 >, D_CYC < 3 >} is in 3-bit binary format. The maximum error from this approximation is obtained when D_STG < 6: 0 > equals 0000111 or 1111111, which leads to an error of 3t_deiay/4.

[0123] The calculation for 3t_deiay/4 is similar to equation (13):

, D_CYC < 1 >}, Equation (14)

[0124] According to equations (13) and (14), the shift-only arithmetic of the PS 915 has an approximation error bounded to 3/4 of the single-stage delay, which leads to negligible impacts on the performance of the EE circuit 875 while greatly simplifying the hardware.

[0125] FIG. 13C illustrates signals associated with the PS 915 in accordance with one or more embodiments. The delay -locking phase is responsible for generating precise pulses with delays of TME/2 or 3 TME/4 to indicate the positive and negative peaks of the output voltage of the ME transducer 125. These pulses serve as control signals for the EE technique. The delay-locking phase is triggered only during the cycles requiring EE for modulating the backscattered magnetic field pulses 825 to save power. During this phase, the pulse generator monitors CNT<4:0> and STG<4:0> to generate peak detection pulses at the rising edge of either the Y<3 :0> or YB<3 :0> signal based on the TDC result and the desired phase shifts. To ensure sufficient amplitude change for high SNR, EE is designed or configured to be performed over four ME cycles, which requires the generation of peak detection pulses in each ME cycle. However, due to the biasing flip in the EE technique, the comparison-based zero-crossing detection can only be reliably used in the first ME cycle. To avoid excessive computing overhead in the pulse generator and the need for a larger- stage RO counter, the delay-locking scheme is used to detect the time of 360° phase shift of the initial zero-crossing detection pulse. At this point, the circuit regenerates the RO start signal such that the delay -locking process may be repeated in the following three ME cycles to be consistent with the first ME cycle. The delay stages in the RO are reset by properly connecting the output to the supply or ground using a short reset signal generated just before the start signal to ensure correct delay counting. Furthermore, the RO employs a 4-stage differential structure to facilitate easier delay computation of TME/4 or 3 TME/4 and lower power consumption.

[0126] FIG. 14 illustrates a low-power, low-latency comparator 940 in accordance with one or more embodiments. The comparator 940 can be used for precise zero-crossing point detection and its performance in optimizing detection delay. A low-power, low-latency comparator 940 can be used for precise zero-crossing point detection and detection delay. Achieving precise zero-crossing detection requires minimizing the delay in the comparator 940. Increasing the biasing current is one approach to reducing the comparator delay, but it comes at the cost of higher power consumption. Furthermore, during the corresponding ringdown phase 820, the ME voltage (z.e., electrical signal pulses 890) gradually decreases resulting in a larger delay in detecting the zero-crossing during later cycles. To address these challenges, an alternative solution that uses an adaptively compensated biasing current to create an unbalanced feature in the comparator 940 is used. The sinking current Ii discharges the internal node A to cause the output voltage to go high earlier when VACI > VAC2. This method significantly reduces the delay in detecting the zero crossing of the ME voltage without consuming a substantial amount of additional power. In later ME cycles with smaller ME voltages, the delay can be overcompensated due to the increased sensitivity to the biasing current Ii. To address this issue, another sourcing current to node A is added, which increases as the number of ME cycles increases. This is achieved by a PMOS-based current source whose gate voltage tracks the decaying envelope of the input voltage. To further improve the reliability of this mechanism, the biasing current can be calibrated through the downlink data to ensure effectiveness. With this mechanism, the delay of zero-crossing detection is significantly reduced and demonstrates consistent performance over the entire corresponding ringdown phase across different process corners. It is worth noting that the zero-crossing detection comparator 940 is only enabled at the desired cycle where amplitude modulation is needed to save power consumption.

[0127] FIG. 15A illustrates a switched-capacitor array 875a in accordance with one or more embodiments. The switched-capacitor array 875a is a type of EE circuit 875. FIG. 15B illustrates corresponding signals 945 in accordance with one or more embodiments. The switched-capacitor array 875a may include a 4-phase switched-capacitor array to efficiently extract stored energy from the ME materials of the ME transducer 125. Through charge redistribution, the charge stored in the piezoelectric capacitor Cp of the ME transducer 125 is transferred to the switched-capacitor array 875a. In each phase, starting from dq and progressing up to <t>₂, the capacitors of the switched-capacitor array 875a are sequentially connected to the ME transducer 125. Through this sequential connection, the electrical energy of the ME transducer 125 rapidly dissipates to substantially reduce the mechanical energy of the piezoelectric material of the ME transducer 125 and attenuate the strength of the backscattered magnetic field pulses 825. To accommodate the alternating nature of the output voltage of the ME transducer 125, two switches are arranged in series to ensure a complete interruption of the load path and the efficacy of power recovery during the normal power transfer phase. Furthermore, the switches are controlled by nonoverlapping short pulses generated by a delay-chain circuit when the voltage of the ME transducer 125 reaches its peak.

[0128] Similar to SSHC, EE employs sequentially activated switched capacitors to extract energy from the piezoelectric-layer capacitor Cp at the peak points of the sinusoidal electrical signal pulses 890, where Cp holds the most energy. Following energy extraction, Cp is used to flip the voltage VME of the ME transducer 125 to effectively boost the amplitude of VME in the next half cycle, if VME is not clamped by the output of the rectifier. When EE boosts the voltage of the ME transducer 125, the output power of the ME transducer 125 (PME = VME x IME) increases, since the ME current remains unchanged due to the high quality factor of the ME transducer 125. In this way, the mechanical energy in the ME transducer 125 can be quickly extracted and dissipated (or transferred to the output of the rectifier if clamped). The voltage flipping process repeats eight times over four ME cycles to ensure sufficient energy extraction and amplitude reduction while preserving a good timing resolution for PWM. Four flying capacitors with a total on-chip capacitance of 1.2 nF are used for the EE circuit 875 considering the trade-offs among hardware complexity, on-chip area, and energy extraction performance.

[0129] FIG. 16 shows a schematic for the AFE 950 in accordance with one or more embodiments. In some embodiments, the system-on-chip 155 of the implant 100 includes the AFE 950 and supports continuous data streaming. To expand the input dynamic range (> 60 dB) and eliminate the need for a high-resolution ADC (ENOB > 15 bits), the first stage of the AFE 950 is implemented as a moderate-gain (gain = 15), current-reusing two- stage LNA. Because of its moderate gain, the neural signal (z.e., user data 850) is less likely to saturate the amplifier. Since the LNA drives the ADC directly, a highly energy efficient third-order noise-shaping SAR ADC 955 is implemented, which features a small sampling capacitor to relax the driving strength requirement of the LNA. The SAR ADC 955 has a 1 kHz bandwidth and an oversampling ratio (OSR) of eight. Due to the low OSR, the subsequent decimation filter is a simple 3 -stage cascaded integrator-comb (CIC) filter. Since the amplitude of the neural signal is relatively small and the frequency of the artifacts is low compared to the bandwidth of the SAR ADC 955, a delta modulator is added to further reduce the final output to a 2 kSa/s 8-bit data stream for wireless uplink transmission.

[0130] FIG. 17 shows time-varying electrical driving signals associated with the EE technique in accordance with one or more embodiments. It is evident that with the assistance of peak detection, the system-on-chip 155 of the implant 100 extracts energy near the maximum voltage of the electrical signal pulses 890. After four cycles of EE, the stored energy in the ME transducer 125 rapidly dissipates to reduce the ME output voltage. The first detection pulse is generated at the positive peak voltage and the subsequent detection pulse exhibits a delay of 3.079 microseconds (ps), which is close to the period of the backscattered magnetic field pulses 825 of 3.021 ps. This demonstrates the effectiveness of the proposed peak detection technique.

[0131] By accurately timing the extraction of the energy of the ME transducer 125 at the implant 100, controlled by the uplink data 865 to be transmitted, the modulated backscattered electrical signal pulses 960 exhibit varying pulse widths 965. This variation is represented by substantial amplitude changes at different time points within the corresponding ringdown phase 820.

[0132] FIG. 18 illustrates the circuit diagram of the external transceiver 110 in accordance with one or more embodiments. The external transceiver 110 may be built from off-the-shelf components to transfer power from a battery 970 of the external transceiver 110, optionally downlink data 866 to the implant 100, and receive and decode the encoded uplink data 871. [0133] As described above, EE quickly dissipates energy in the ME transducer 125 and induces an abrupt amplitude reduction in the backscattered magnetic field pulses 825 for PWM. Compared to FSK modulation in the time-domain, PWM improves both SNR and data rate because it avoids deviating from the resonant frequency of the ME transducer 125 and transmits three or more bits in a single corresponding ringdown phase 820 to amortize the excitation phase time and corresponding ringdown phase time.

[0134] To better analyze the PWM technique, a high-order ME impedance model 975 for ME transducers 125 is also disclosed as illustrated in FIG. 19. The high-order ME impedance model 975 accounts for the high-frequency resonance modes of the ME transducer 125 by attaining excellent matching between simulation and measurement.

[0135] While using a printed circuit board (PCB)-level EE prototype to validate the concept of the high-order ME impedance model 975, noticeable deviations from the simulation results obtained with a basic ME model are observed. After voltage flipping, the voltage amplitude of the ME transducer 125 first decreases and then increases in the opposite direction. Moreover, frequency components that are higher than the carrier frequency appear after voltage flipping and persist several more cycles after completing the EE operations. The peak points become difficult to distinguish and the residual voltage after EE is higher than the baseline.

[0136] Prior works have also revealed the second vibration mode of an ME transducer 125 and utilized it for uplink communication. However, there is no ME impedance model in the literature that describes the other vibration modes. Here, the disclosed high-order ME impedance model 975 captures high-frequency vibration modes in EE, which can potentially be extended to other designs.

[0137] The primary vibration mode of an ME film of the ME transducer 125 is the length mode, which operates at a low frequency of hundreds of kilohertz. However, the ME film also exhibits piezoelectric effects in other directions, such as width and thickness. These modes have higher frequencies and generally contribute minimally to voltage of the ME transducer 125 when driven by a magnetic field, as they are far off-resonance. During EE, however, the EE circuit 875 induces sudden drops in voltage of the ME transducer 125, which not only extracts energy from the main resonance mode but also excites other resonance modes due to the wide bandwidth of the pulse-shaped signals. Consequently, these extra resonance modes affect the EE process.

[0138] In the high-order ME impedance model 975, each resonance is represented by an impedance Z_m, where the R_m, L_m, and C_m correspond to damping, mass, and stiffness in the mechanical domain. C_p is the lumped electrical parasitic capacitor between the two terminal s of the ME transducer 125. Using the high-order ME impedance model 975, the behavior of the ME transducer 125 can be analyzed in EE through simulation tools. Note that the high- order ME impedance model 975 excludes input sources in each branch, as the external transmitter 165a is turned off during EE. However, if the behavior needs to be examined when the ME film is powered by a magnetic field, the sinusoidal source and transformer can be introduced to the main resonance branch while treating the other branches as loads.

[0139] FIGs. 20A and 20B describes a method and illustrates a PWM system 2000 that relies on PWM technique in accordance with one or more embodiments. By performing the following described method, uplink data recorded by an implant 2002 may be wirelessly transmitted to an external transceiver 2004 using backscattered field pulses 2006 induced by a transducer 2008 of the implant 2002. It is to be understood that one or more of the blocks shown in the flowchart may be omitted, repeated, and/or performed in a different order than the order shown. Further, one or more of the blocks may be performed substantially in series or in parallel (z.e., simultaneously). Accordingly, the scope of the disclosure should not be considered limited to the specific arrangement of blocks shown in the flowchart.

[0140] In block 2009, an external transceiver 2004 induces field pulses 2010. To do so, an external controller 2012 of the external transceiver 2004 may prescribe an amplitude and phase of a current to flow through an external transducer 2014 such that the external transducer 2014 emits the field pulses 2010.

[0141] In block 2016, the transducer 2008 of the implant 2002 receives the field pulses 2010 as shown by the arrow 2018a. Accordingly, the field pulses 2010 propagate the predetermined distance 2020 between the external transceiver 2004 and the implant 2002.

[0142] In block 2022, the transducer 2008 converts the field pulses 2010 to electrical signal pulses. [0143] In block 2024, the transducer 2008 induces the backscattered field pulses 2006 based on the field pulses 2010.

[0144] In block 2026, a sensor (not shown in FIG. 20B) of the implant 2002 senses and records the uplink data. The uplink data may be associated with the implant 2002 and/or a user (not shown in FIG. 20B) that the implant 2002 is disposed or implanted within.

[0145] In block 2028, encoded uplink data is determined by encoding, using an uplink transmitter (not shown in FIG. 20B) of the implant 2002, the uplink data to align with the corresponding ringdown phase of each of the backscattered field pulses 2006.

[0146] In block 2030, the transducer 2008 induces modulated backscattered field pulses by affecting the backscattered field pulses 2006 within the corresponding ringdown phase based on the encoded uplink data.

[0147] In block 2032, an external receiver 2034 of the external transceiver 2004 receives the modulated backscattered field pulses as shown by arrow 2018b. Accordingly, the modulated backscattered field pulses propagate the predetermined distance 2020 between the external transceiver 2004 and the implant 2002.

[0148] In block 2036, the external receiver 2034 converts the modulated backscattered field pulses to modulated backscattered electrical signal pulses. The modulated backscattered electrical signal pulses include the encoded uplink data.

[0149] In block 2038, the external controller 2012 of the external transceiver 2004 determines the uplink data by decoding the encoded uplink data from the modulated backscattered electrical signal pulses.

EXAMPLES

[0150] The below illustrates examples of an AE system 130 configured to power an implant 100 no matter the alignment or misalignment of the implant 100 relative to the external transceiver 110. However, any part, specification, step, etc. described in association with the AE technique or system configured to perform the AE technique may be used in association with the PWM technique or system configured to perform the PWM technique and vice versa without departing from the scope of the disclosure. [0151] The implant 100 includes a fully-integrated 14.2 mm³ implantable stimulator embedding a custom low-power system-on-chip 155 powered by an ME film as the transducer, an external transceiver 110 with a custom three-channel external receiver chip 170, and an external transmitter coil array 115 with mutual inductance cancellation. The external receiver chip 170 achieves -161 dBm/Hz input-referred noise with 64 dB gain tuning range to reliably sense the first magnetic field emitted from the implant coil 120 of the implant 100. The external receiver chip 170 also offers fast phase detection for driver control.

[0152] The system-on-chip 155 of the implant 100 and three-channel external receiver chip 170 are fabricated in TSMC 180 nm CMOS technology with respective areas of 1.34 and 1.62 mm². The performance of the channels of the external receiver chip 170 are tested and integrated into the complete wireless AE system 130 for evaluation.

[0153] When a low-impedance LC loading is presented to the differential output of the external transmitter driver 165, the voltage swing decreases, and the LC loading reshapes the output time-series. To better assess the output time-series and spectrum of the three- level PA, the LC loading at the external transmitter driver 165 output is removed to obtain the measurement results). The external transmitter driver 165 is powered by a 1 V LDO separately to ensure better isolation from other circuits during operation. Here, the clock duty cycle produced by DCC is set to around 30%. As designed, the output voltage briefly stays near half VDD when all transistors of the PA are off. The differential output of the AE TX is a symmetrical three-level time-series that better mimics a sinusoidal signal than a basic square signal. By adjusting the biasing voltage of the DCC block, the duty cycle is tuned to achieve better harmonic suppression. During testing, the output spectrum achieves an approximately 30 dB third-order harmonic reduction with only a 2.2 dB fundamental tone loss at around 30% duty cycle. Since the external transmitter driver 165 operates for only a few cycles after activation, its power consumption is greatly amortized, and the average power remains small. The power consumption linearly increases with AE activation frequency, where the multilevel technique reduces power by 27% at 100 Hz. By extending the regression lines to 0 Hz, both two and three-level PAs have a similar static power of roughly 50 nW. The gain and bandwidth of a single AFE channel are measured with an impedance analyzer. LNA achieves a 22 dB tuning range by configuring the resistor within the TIA stage. The first-stage PGA has the same gain control strategy with a 37 dB tuning range. To compensate for PVT variations and balance the three channels, PGA2 offers a 5 dB tuning range with less than 1 dB step. High-pass cutoff frequency has four configurations from 100 to 600 kHz to filter low-frequency interference and noise. Across various configurations of the AFE, measurements closely match the simulations. The slight mismatch at high gains is due to the limited SNR of the signal source.

[0154] For system demonstrations, the channel gain is set to ~63 dB at 1.5 MHz to achieve the largest VR_X swing (around 1.5 V) with perfectly aligned external transmitter coil array 115 and transducer. The silicon mismatch after calibration is less than 0.2 dB. The fluctuation is caused by limited equipment calibration accuracy. The input-referred noise is around -161 dBm/Hz, calculated from the output noise floor and the channel gain.

[0155] Next, the integrated PD and ADC are tested. A 1 MHz sinusoidal signal is input into the peak detector, which converts it to a DC value. The DC value is then digitized by a cascaded SS-ADC using a 400 MHz external clock. The output digital code is linear when the input voltage sweeps from 0.1 to 1.2 V. The R² value of the transfer curve is 0.996. The linearity is primarily restricted by the gain error variations of OpAmps in the ramp signal generator and the peak detector. The linearity of the peak detector is the more dominant factor since the speed of the ramp signal is much lower than the desired electrical signal.

[0156] The power measured from the single channel external receiver chip 170 is 18.65 mW at a 1.8 V supply voltage, including 10.2 mW LNA power, 3.5 mW PGA power, 2.5 mW ADC power, and 2.45 mW from clock path and digital circuity. The LNA consumes more than half of the total power to reduce input-referred noise for weak AE signal sensing, which is overdesigned to guarantee a sufficient SNR when the implant 100 is positioned a little further away from the external transmitter coil array 115.

[0157] Off-the-shelf coils made with Litz wire are assembled into the external transmitter coil array 115 that includes three external coils 105a. Each external coil 105a has its own compensation capacitor to match the resonant frequency of the ME film (around 340 kHz) for high PTE. The placement of the three external coils 105a reduces the mutual inductance. When the center-to-center distance of the external coils 105a is 24 mm, the external coils 105a have minimum coupling. The abrupt phase change at this point also indicates that the amplitude of S21 crosses zero. The measurement also closely matches the simulation results.

[0158] In vitro experiments with the complete system enabled by the two chips are performed. Three off-the-shelf GaN half-bridge drivers (MASTERGAN 1 from STM) were used to drive the TX LC tanks efficiently. The half-bridge configuration ensures a grounded connection during the AE sensing phase. The implant chip packaged with an ME film and an AE coil was installed on a motor stage capable of moving in three directions and rotating in an X Z plane via a servo motor. A 3-D-printed plastic rotation bar extends from the rotation axis of the servo motor to mount the implant under test. A wood bar connector linked the implant’s rotation motor to the connecting point of the motion system, minimizing interference from the motion system. A 1.5-cm porcine tissue was placed between three TX coils and the implant board. On the RX side, the RX chip was connected to the output of the ME driver through an isolation capacitor. This isolation capacitor creates a high pass comer with the input impedance of the AE RX to remove the de draft of the GaN driver’s output. Additionally, a protection switch was inserted at the beginning of the RX chain and controlled by NI DIO, synchronizing with the TX driver. The DIO collected the three digitized amplitudes and three polarities from the RX chip and provided the desired PWM control signal to the drivers.

[0159] The AE system 130 cycles through charging, downlink programming, AE sensing, and stimulation. Here, the biphasic stimulator is programed to 3 V 0.4 ms pulsewidth with 10 and 20 ms intervals.

[0160] The closed-loop WPT performance across various implant positions and orientations in in vitro condition are evaluated. In different cases, the amplitude and phase detection were working properly, and the smaller ADC output value indicated a larger amplitude, which could be concluded first as the reference clock of phase detection. When the coupling coefficient of one channel is significantly lower than another, falling below the empirical threshold of 8, the driver is deactivated. This is because PWM control lacks precision under these conditions, and the static power consumption of the channel is relatively high, while its contribution to magnetic field generation is minor.

[0161] To confirm that the AE solution is indeed optimal, the current ratio in external transmitter coil array 115 is manually swept in two representative cases (2 and 4). For simplicity, first external coil 105a is turned off in this test and normalize the PTE to the peak point in each case. In case 2, the special arrangement allows for a simple theoretical calculation of the optimal ratio between the second and third external coil 105a, which is 1. The manual sweeping test agrees with the calculation. The AE solution is merely 1.8% away from the PTE achieved by ideal control. In case 4, the implant 100 has an arbitrary set of lateral and angular/rotational offsets. Guided by theoretical analysis, finer sweeping steps are taken to find the optimal condition in the testing. Again, the AE scheme achieves the expected result and significantly increases the efficiency of power transfer by 66% with only a 2% drop from the highest PTE found by manual sweeping. This tracking error is mainly incurred by the nonlinearity of the external receiver chip chain, because an openloop structure is used for the front end to reduce noise and power.

[0162] The PTE measurements demonstrate how AE compares with three baselines in terms of sensitivity to rotational and lateral misalignment. The baselines include: 1) a single coil in our optimized three-coil array placed at the origin (diameter = 4.2 cm); 2) a single coil with roughly the same footprint of three coils combined (diameter = 6.8 cm); and 3) the same three coils as AE but with constant nonadaptive driving currents. The coils are reused for these baselines with the intention of reducing the impacts of coil designs (parameters and number of turns) on field distribution and PTE measurements. It can be observed that three coils, even without adaptation, outperform the single-coil baseline with consistently higher PTE over both lateral and rotational mismatch. AE offers superior PTE across all tested poses, the improvements are more significant under severe misalignment. Compared to the single-coil baseline, AE improves the PTE by 6.8* with 90° rotation from the ideal orientation and by 18x with 20 mm lateral offset, a type of translational offset.

[0163] The AE technique enables the first-in-class omnidirectional WPT system, maintaining more than half of the perfect alignment PTE across all rotations angles. Accordingly, using an auxiliary coil on the implant to adaptively control a multicoil TX array for optimal magnetic flux steering to overcome the primary challenges of coupling sensing and adaptive multicoil WPT with weakly coupled TX and RX is feasible. A complete bioelectronic implant system is demonstrated, including a 14.2 mm3 implantable neurostimulator with a custom SoC powered by a ME film, an external TX including a custom AE RX chip, and a mutual inductance canceled multicoil TX array. Two custom chips with various circuit techniques are built to efficiently and reliably perform AE signal transmission and reception. The prototyping system functions properly in a fully wireless in vitro testing setup and demonstrates omnidirectional power transfer capabilities, showcasing a 6.8x higher PTE than a single-coil baseline under 90° rotation from the ideal alignment, as well as less than 2% efficiency drop compared to using ideal control.

[0164] FIG. 21A and FIG. 21B display PTE under rotational and lateral offsets. The AE technique that relies on three external coils 105a, no AE technique that relies on three external coils 105a with identical power and phase, and a single external coil 105a at the origin are compared. Using three external coils 105a consistently outperforms the single external coil 105a. Further, the AE technique shows extra enhancements under misalignment due to rotational offset and lateral offset. Particularly, the AE system 130 enhances PTE by 6.8x after a 90° rotational offset from the ideal alignment compared to one external coil 105a, proving the omnidirectional WPT capability of the AE system 130. Further, the tracking error of the AE technique negligibly degrades the PTE by less than 2% relative to ideal control.

[0165] Considering the voltage compliance of the logic-rule system-on-chip 155 of the implant 100, the total power of the external transmitter coil array 115 is kept under 0.5W, which is well below the driver power allowed by IEEE Std (z.e., 17.6W at 340 kHz). At 90° rotational offset, the implant 100 still receives 4.45mW which is sufficient for its operation.

[0166] The below illustrates examples of a PWM system 900 configured to transmit uplink data from the implant 100 to the external transceiver 110.

[0167] The PWM system 900 includes a 6.7 mm³ ME implant and a custom portable external transceiver 110 for power transfer, stimulation, wireless recording, and bidirectional communication. The system-on-chip 155 of the implant 100 uses TSMC 180 nm CMOS technology to integrate the power manager 180, bidirectional data transmission, bio-stimulation, temperature sensing, and LFP/ECoG/EGM/ECG recording, targeting applications in cortical and cardiac implants. Overall, the PWM system 900 demonstrates (1) wireless power, downlink, and PWM backscattering uplink using a single 5x2x0.2 mm³ ME transducer 125; (2) EE attenuating the backscattered magnetic field pulses 825 by more than 50% in 2 ME cycles; (3) PWM ME backscatter achieving 17.73 kbps data rate and 0.9 pj/bit efficiency at 331 kHz carrier; and (4) reliable operation at up to 5 cm distance with less than 8.5x10 ⁵ bit-error-rate (BER) using a lightweight multi-layer-perceptron (MLP) neural network for demodulation.

[0168] The functionalities of the PWM system 900 were evaluated ex vivo by placing a 2 cm -thick porcine tissue between the implant 100 and the external transceiver 110. The external transmitter 165a was programmed to drive the external transmitter coil 805 with a current of a predetermined frequency to transmit power and downlink data 866 to the implant 100 via the magnetic field pulses 800. The external receiver 170a received the modulated backscattered magnetic field pulses and decoded the encoded uplink data 871. The external transmitter coil 805 and external receiver coil 810 are disposed concentrically one within the other as illustrated in FIG. 9 to cancel the transmitter interference. A DC biasing magnet required for ME operation is placed near the implant 100. FIG. 17 displays operating time-varying electrical signals of the implant 100. Specifically, the implant 100 is magnetoelectrically powered and programmed with time-domain modulated magnetic field pulses that include encoded downlink data. After programming, the system-on-chip 155 performs different tasks that includes biphasic stimulation and uplink telemetry. The zoom-in views of FIG. 17 display the voltage of the ME transducer 125 (z.e., modulated backscattered electrical signal pulses 960) during the PWM backscatter uplink. The EE circuit in the form of a switched-capacitor array extracts the energy of the ME transducer 125 at different time points during the corresponding ringdown phase 820 based on first electrical signal. The implant 100 demonstrates energy extraction from the ME transducer 125 near the optimal peak points, which causes an immediate biasing flip. Despite the distortion in the signal, the 4-cycle EE quickly dissipates the energy in the ME transducer 125 to decrease the output voltage of the ME transducer 125 by more than 1 V amplitude. The first peak detection pulse shows an error of 18 ns (z.e., 0.6% of TME), which is sufficient accuracy of the PD.

[0169] The data rate and SNR of the modulated backscattered electrical signal pulses 960 was also evaluated. FIG. 17 displays modulated backscattered electrical signal pulses 960 after AFE with clearly distinguishable pulse widths 965 during the corresponding ringdown phase 820 modulated by the 3-bit encoded uplink data 871. Here, each 3-bit encoded uplink data 871 takes a 24-ME-cycle excitation phase 815, succeeded by a 32- ME-cycle corresponding ringdown phase 820. This results in a data rate of 17.73 kbps with a 331 kHz carrier. To trade-off between the SNR and the distinction between codes, 3- ME-cycle time interval is selected for PWM.

[0170] In this setup, the SNR is defined as the voltage drop in the modulated backscattered electrical signal pulses 960 divided by the noise. Owing to fast amplitude reductions at accurate time intervals, the received modulated backscattered electrical signal pulses 960 achieves an SNR greater than 10.9 dB at up to a predetermined distance of 5 cm. The SNR varies with different uplink data 865 because of the PWM encoding methods. For example, the encoded uplink data pattern “001” is modulated at the beginning of the corresponding ringdown phase 820, where the ME transducer 125 has the highest remaining voltage and thus the highest voltage drop and SNR.

[0171] The encoded uplink data 871 may be decoded using a time-domain drop detection or compact 4-layer multi-layer perceptron (MLP) neural network (hereinafter also simply “MLP”). The modulated backscattered electrical signal pulses 960 after AFE is sampled by a 12-bit ADC at 2 MS/s.

[0172] In the time-domain drop detection method, the envelope of the amplified modulated backscattered electrical signal pulses 960 is extracted by identifying the peaks in the time-series. Next, the voltage differences between adjacent peaks are compared with an adaptive threshold that decays over time. When the AV signal exceeds the threshold, it triggers the decoder to read the corresponding timestamp and generate the output data (z.e., uplink data 865) based on a look-up table (LUT). This method is straightforward and features low latency and low power consumption. However, as the predetermined distance 145 between the implant 100 and external transceiver 110 becomes larger, the time-series is susceptible to noise, interference, and body movement, which causes a large variation in the AV signal and eventually reduces the decoding accuracy.

[0173] To enhance robustness against low SNR signals, a 4-layer MLP using PyTorch is developed. The MLP comprises four layers with 300, 100, 25, and 8 fully-connected neurons, respectively, using ReLU as the activation function. All weights and activations are quantized to 5 bits. The training dataset includes training modulated backscattered electrical signal pulses acquired when the predetermined distance 145 is between 1 cm and 5 cm. Each training sample among the training dataset is segmented by the trigger signal and consists of a 150-ps time-series with 300 data points. During inference, the time- domain training modulated backscattered electrical signal pulses is cropped based on the data packet and processed by the MLP, which classifies the time-series into one of eight output codes. The quantized model uses only 20 KB of memory, making it feasible for inference on an MCU with low cost and low power consumption.

[0174] The drop detection provides sufficient BER (less than 1 x 10 ⁵) within a shorter predetermined distance 145 while the MLP decoding offers higher accuracy across a longer predetermined distance 145 to yield a BER of 8.5x I O ⁵ at a predetermined distance 145 of 5 cm.

[0175] The wireless recording functionality of the PWM system 900 is demonstrated in two stages. First, pre-recorded local field potential (LFP) time-series from rats that mimics uplink data 865 are fed to the implant 100 through wires, then the pre-recorded LFP time-series are wirelessly transmitted to the external receiver 170a using the PWM technique. The modulated backscattered electrical signal pulses 960 closely match the ground truth. Second, an encapsulated untethered implant 100 is tested in vitro by submerging the implant 100 in phosphate-buffered saline (PBS). The external transmitter coil 805 and external receiver coil 810 are placed under the beaker that contains the PBS for power transfer and communication. The pre-recorded LFP time-series are transmitted through a pair of electrodes immersed in PBS. The working electrode is placed close to the implant 100, while the reference electrode is placed further away on the other side so that the system-on-chip 155 can receive the differential signal.

[0176] The system-on-chip 155 consumes a total power of 74.8 pW with a power transfer efficiency of 0.37%. About 80% of this power is used by the power manager 180 and control blocks. The uplink transmitter 870 consumes a minimal power of 0.45 pW, which accounts for only 0.6% of the total power, thanks to the PWM technique and highly duty-cycled circuit design. The uplink communication bit energy, defined as the uplink power divided by data rate, is 0.9 pj/bit. The low-bit energy is achieved thanks to the highly duty-cycled peak detector, which consumes the most power of the uplink transmitter 870. The bit energy remains unchanged when the predetermined distance 145 varies because the uplink power is insensitive to the predetermined distance 145.

[0177] SNR testing was performed at the output of the AFE of the external transceiver 110 considering both environmental noise and circuit noise. As different 3-bit data sequences involve EE at different ME cycles during the corresponding ringdown phase 820, they exhibit varying voltage drops and, consequently, different SNR values. Therefore, the SNR for uplink data from ”001” to ”111” is measured. The uplink data ”111,” with EE occurring in the final few ME cycles, exhibits the smallest amplitude change. However, even in this scenario, the SNR remains at a minimum of 16.73 dB at a 3-cm communication distance, which ensures reliable data recovery.

[0178] In summary, the experiment involves testing an implant 100 positioned 2 cm away from the external transceiver 110. Each 3 -bit data transmission cycle consists of a 24- ME-cycle excitation. The implant 100 consumes an average power of 15.6 nW in each data transmission phase, leading to low energy consumption of 0.9 pj/bit. The uplink data 865 is readout from the testing pin of the implant 100, revealing how the 3 -bit data corresponds to different time points of amplitude reduction in the backscattered signal. The modulated backscattered electrical signal pulse 960 is measured at the output of the AFE of the external transceiver 110, which utilizes an amplification gain of 52 dB.

[0179] Although only a few example embodiments have been described in detail above, those skilled in the art will readily appreciate that many modifications are possible in the example embodiments without materially departing from this invention. Accordingly, all such modifications are intended to be included within the scope of this disclosure as defined in the following claims.

Claims

CLAIMS What is claimed:

1. A system comprising: an implant comprising a first implant transducer configured to induce a first field; and an external transceiver disposed up to a predetermined distance away from the implant, wherein an alignment of the implant relative to the external transceiver is unknown, the external transceiver comprising: an external transmitter transducer array comprising at least two external transducers, each of the at least two external transducers configured to: receive the first field; and convert the first field to a first electrical signal, and an external controller electrically coupled to the external transmitter transducer array and configured to: for each of the at least two external transducers: receive the first electrical signal, and determine a received amplitude of the first electrical signal and a received phase of the first electrical signal; determine a first transmitted amplitude of a first electrical driving signal and a first transmitted phase of the first electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers, and determine a second transmitted amplitude of a second electrical driving signal and a second transmitted phase of the second electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers, wherein the first transmitted phase and the second transmitted phase are a same phase or an opposing phase, wherein a first external transducer among the at least two external transducers is configured to induce, using the first electrical driving signal, a second field, and wherein a second external transducer among the at least two external transducers is configured to induce, using the second electrical driving signal, a third field, wherein the implant further comprises a second implant transducer aligned relative to the first implant transducer, the second implant transducer configured to: receive the second field and the third field, and convert the second field and the third field to a second electrical signal, wherein the second electrical signal is configured to power the implant.

2. The system of claim 1, wherein the at least two external transducers are disposed within planes parallel to one another and overlap with one another.

3. The system of claim 1, wherein the external transmitter transducer array comprises a third external transducer, and wherein the first external transducer, the second external transducer, and the third external transducer are disposed within planes parallel to one another and do not overlap with one another.

4. The system of any one of claims 1-3, wherein the alignment comprises a misalignment comprising at least one of a rotational offset or translational offset.

5. The system of any one of claims 1-4, wherein the first implant transducer comprises an implant coil, and wherein an outer diameter of the implant coil is one centimeter or less.

6. The system of any one of claims 1-5, wherein the at least two external transducers comprises at least two external coils, and wherein an outer diameter of each of the at least two external coils is at least two centimeters, inclusive.

7. The system of any one of claims 1-6, wherein the predetermined distance is up to ten centimeters, inclusive.

8. The system of any one of claims 1-7, wherein the second implant transducer comprises a magnetoelectric (ME) transducer, and wherein the ME transducer is aligned perpendicular to the first implant transducer.

9. The system of any one of claims 1-8, wherein the implant further comprises an energy extraction circuit, and wherein the energy extraction circuit comprises a switched- inductor or switched-capacitor array.

10. A method comprising: inducing, using a first implant transducer of an implant, a first field, wherein the implant is disposed up to a predetermined distance away from an external transceiver, and wherein an alignment of the implant relative to the external transceiver is unknown; receiving, by an external transmitter transducer array of the external transceiver, the first field, wherein the external transmitter transducer array comprises at least two external transducers; for each of the at least two external transducers: converting the first field to a first electrical signal, and determining a received amplitude of the first electrical signal and a received phase of the first electrical signal; determining a first transmitted amplitude of a first electrical driving signal and a first transmitted phase of the first electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers; determining a second transmitted amplitude of a second electrical driving signal and a second transmitted phase of the second electrical driving signal using the received amplitude and the received phase independently for the at least two external transducers, wherein the first transmitted phase and the second transmitted phase are a same phase or an opposing phase; inducing a second field by applying the first transmitted amplitude at the first transmitted phase to a first external transducer among the at least two external transducers; inducing a third field by applying the second transmitted amplitude at the second transmitted phase to a second external transducer among the at least two external transducers; receiving, by a second implant transducer of the implant, the second field and the third field, wherein the second implant transducer and the first implant transducer are aligned; converting, using the second implant transducer, the second field and the third field to a second electrical signal; and powering the implant using the second electrical signal.

11. The method of claim 10, wherein receiving the first field occurs over a first time interval, and wherein inducing the second field and inducing the third field occurs over a second time interval.

12. The method of claims 10 or 11, wherein the second field comprises second field pulses, wherein the third field comprises third field pulses, wherein the second electrical signal comprises second electrical signal pulses, and wherein powering the implant comprises intermittently powering the implant using the second electrical signal pulses.

13. A method comprising: inducing, using an external transceiver, field pulses, wherein each of the field pulses has an excitation phase and a corresponding ringdown phase; receiving, by a transducer of an implant, the field pulses, wherein the implant is disposed up to a predetermined distance away from the external transceiver; converting, using the transducer, the field pulses to electrical signal pulses; inducing, using the transducer, backscattered field pulses based on the field pulses, wherein each of the backscattered field pulses has the excitation phase and the corresponding ringdown phase; sensing and recording uplink data; determining encoded uplink data by encoding the uplink data to align within the corresponding ringdown phase; inducing, using the transducer, modulated backscattered field pulses by affecting the backscattered field pulses within the corresponding ringdown phase based on the encoded uplink data, wherein the corresponding ringdown phase of each of the modulated backscattered field pulses comprises at least three bits of the encoded uplink data; receiving, by the external transceiver, the modulated backscattered field pulses; converting the modulated backscattered field pulses to modulated backscattered electrical signal pulses comprising the encoded uplink data; and determining the uplink data by decoding the encoded uplink data from the modulated backscattered electrical signal pulses.

14. The method of claim 13, further comprising: intermittently powering the implant during the excitation phase using the electrical signal pulses.

15. The method of claim 13 or 14, wherein encoding the uplink data comprises: determining a start of the corresponding ringdown phase based on the electrical signal pulses; and determining a peak of a voltage of the electrical signal pulses.