JP2012106080A - Cleaning device equipped with debris sensor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an autonomous type cleaning device, capable of controlling an action mode, a moving pattern, or an action to immediately detect collision of debris on a floor surface as a cleaning object, and respond to detected debris to move the device to a more contaminated area.SOLUTION: By means of a piezoelectric debris sensor which responds to the collision of debris, and related signal processors, the autonomous or non-autonomous cleaning device detects presence of debris, and responds to select an action mode, an action condition, or moving pattern such as spot coverage. A plurality of sensor channels are usable to enable detection or creation of differential right/left debris signals, so that the autonomous device is moved toward the debris. By positioning the debris sensor to a place where debris swept up by a brush assembly is colliding with each other, collision of the debris can be more properly detected.

Description

This patent application is related to the following commonly owned US patent applications or patents, the contents of which are hereby incorporated by reference as if fully set forth.

US patent application Ser. No. 09 / 768,773, filed Jan. 24, 2001, now US Pat. No. 6,594,844, entitled “Robot Obstacle Detection System (Atty. Dkt. . DP-4) "
US Provisional Patent Application No. 60 / 345,764, filed Jan. 3, 2002, titled “Cleaning Mechanisms for Autonomous Robots”
US patent application Ser. No. 10 / 056,804, filed Jan. 24, 2002, entitled “Method and System for Robot Localization and Confinement (Atty Dkt. DP. DP-6) "
US patent application Ser. No. 10 / 167,851, filed Jun. 12, 2002, entitled “Method and System for Multi-Mode® Coverage for Method and System for Autonomous Robots” an Autonomous Robot (Atty Dkt. DP-5) "
US patent application Ser. No. 10 / 320,729, filed Dec. 16, 2002, titled “Autonomous Floor-Cleaning Robot (Atty Dkt. DP-10)”
US patent application Ser. No. 10 / 661,835, filed Sep. 12, 2003, titled “Navigational Control System for Robotic Device (Atty Dkt. DP-9)”
The present invention relates generally to cleaning devices, and more particularly to a debris sensor for sensing instantaneous impacts due to debris in the cleaning path of the cleaning device and for enabling control of the operating mode of the cleaning device. The term “debris” is used herein to collectively indicate dirt, dust, and / or other particulates or objects collected by a vacuum cleaner or other cleaning device, whether autonomous or non-autonomous. Used for.

BACKGROUND OF THE INVENTION Debris sensors, including some suitable for cleaning devices, are known in the art. The debris sensor is beneficial in autonomous and non-autonomous cleaning devices, such as disclosed in the above-referenced patent applications, and whether the user is entering a particularly dirty area. To increase the power setting in response to debris detection or to allow some other operational settings to be modified.

  Examples of debris sensors are disclosed below.

De Brey 3,674,316
De Brey 3,989,311
De Brey 4,175,892
Kurz 4,6011,082
Westergren 4,733,430
Martin 4,733,431
Harkonen 4,829,626
Takashima 5,105,502
Takashima 5,136,750
Kawakami 5,163,202
Yang 5,319,827
Kim 5,440,216
Gordon 5,608,944
Imamura 5,815,884
Imamura 6,023,814
Kasper 6,446,302
Gordon 6,571,422
Of the examples disclosed herein, many such debris sensors are optical in nature and use light emitters and detectors. For example, in a typical design used in a vacuum cleaner, the optical sensor's optical transmitter and optical receiver are positioned so that they are exposed to a suction or clean path through which dust flows. Thus, during use of the vacuum cleaner, dust particles tend to adhere to the exposed surfaces of the optical transmitter and optical receiver where light is emitted and detected, degrading the performance of the optical sensor.

  Therefore, it is desired to provide a debris sensor that does not deteriorate due to debris adhesion.

  In addition, debris sensors typical of the prior art are sensitive to the level of debris deposited in the accumulator or clean path, but are not particularly sensitive to instantaneous debris impacts or encounters.

  Thus, debris collisions can be sensed and responded instantaneously, responding immediately to debris on the floor or other surface to be cleaned, against atmospheric flow fluctuations, instantaneous power, or other operating conditions of the cleaning device It would be desirable to provide a debris sensor with reduced sensitivity.

  Also, an autonomous clean having an operating mode, movement pattern, or behavior in response to detected debris, for example, maneuvering the device towards a “dirty” area based on the signal generated by the debris sensor It is also beneficial to provide a device.

  Furthermore, it would be desirable to provide a debris sensor that can be used to control, select or change the mode of operation of either an autonomous or non-autonomous cleaning device.

  The present invention provides a debris sensor and a device that utilizes such a debris sensor, wherein the sensor responds instantaneously to a debris impact and the mode of operation of an autonomous or non-autonomous cleaning device that includes such a sensor. Can be used to control, select, or change.

  One aspect of the present invention is an autonomous cleaning device, a drive system operable to allow movement of the cleaning device, and a controller in communication with the drive system, the control system controlling the drive system. A controller including a processor operable to provide at least one movement pattern of the cleaning device, a brush assembly operable to sweep debris toward a cleaning path of the cleaning device, and the cleaning device including the debris. And a debris sensor that generates a debris signal indicating that the device has encountered a response, wherein the processor selects one operating mode from a predetermined operating mode of the cleaning device in response to the debris signal.

  The selection of the operating mode can include selecting a movement pattern of the cleaning device.

  The movement pattern can include maneuvering the cleaning device toward spot coverage of an area including debris or an area including debris. The debris sensor may include equally spaced first and second debris sensing elements operable to generate first and second debris signals, respectively, and the processor may include the first and second debris signals. In response, each of the movement patterns can be selected to steer towards the more debris side (eg, left or right).

  A debris sensor is located near the cleaning path of the cleaning device where the debris swept up by the brush assembly collides, and in response to hitting the debris, a piezoelectric that generates a signal indicating that it has hit the debris. Sensor elements can be included.

  The debris sensor of the present invention can also be incorporated into a non-autonomous cleaning device. This aspect of the present invention is located in proximity to the clean path and in response to impacting debris, a piezoelectric sensor element that generates a debris signal indicating that it has impacted debris, and in response to the debris signal, A processor may be included to change the operating mode of the cleaning device. Changing the operating mode can be done by turning on a user-recognizable indicator light, changing the power setting (for example, higher power setting when more debris is encountered), or reducing the movement speed of the device. Can be included.

  A further aspect of the present invention is a debris sensor, located near or within a cleaning path of a cleaning device, that generates a first signal indicating collision with debris in response to collision with debris. And a processor operable to process the first signal and generate a second signal indicative of the characteristics of the debris encountered by the cleaning device. The feature may be, for example, a vector from a debris amount or volume parameter, or a current location of the cleaning device to an area containing debris.

  Another aspect of the present invention utilizes the motion of an autonomous cleaning device that covers a floor or other surface to process the debris signal and calculate the debris gradient in conjunction with knowledge about the movement of the cleaning device. The debris gradient indicates the change in debris collision count as the autonomous cleaning device moves along the surface. By examining the sign of the slope (positive or negative in relation to increasing or decreasing debris), the autonomous cleaning device controller continuously adjusts the device path or movement pattern to effectively debris the region. Clean.

  These and other aspects, features, and advantages of the present invention will become more apparent from the following description with reference to the accompanying drawings, in which embodiments of the invention are shown and described by way of illustration.

It is a top view of the typical autonomous cleaning apparatus as an example which can employ | adopt the debris sensor of this invention. FIG. 2 is a block diagram of exemplary hardware elements of the robotic device of FIG. 1 and includes the debris sensor subsystem of the present invention. FIG. 2 is a side view of the robot apparatus of FIG. 1 showing a debris sensor according to the present invention located in a clean or vacuum path where debris swept by a main clean brush element collides. It is an exploded view of the piezoelectric debris sensor according to the present invention. 1 is a schematic diagram of a debris sensor signal processing architecture according to the present invention. FIG. FIG. 6 is a schematic diagram of a signal processing circuit for the debris sensor architecture of FIG. 5. FIG. 6 is a schematic diagram of a signal processing circuit for the debris sensor architecture of FIG. 5. FIG. 6 is a schematic diagram of a signal processing circuit for the debris sensor architecture of FIG. 5. It is a schematic diagram which shows the debris sensor in a non-autonomous cleaning apparatus. 4 is a flowchart of a method according to an embodiment of the present invention.

A more complete understanding of the present invention and the attendant features and advantages will be obtained by reference to the following detailed description of the invention, taken in conjunction with the accompanying drawings.
DETAILED DESCRIPTION OF THE INVENTION Although the debris sensor of the present invention can be incorporated into a wide range of autonomous cleaning devices (and in fact also in the non-autonomous cleaning device shown as an example in FIG. 7), first, FIG. This is described in the flow of a typical autonomous cleaning device shown in FIG. Further details of the structure, function and mode of action of such an autonomous cleaning device are described in the patent applications cited in the cross-reference section above, each of which is incorporated herein by reference. To do. Therefore, the following detailed description is composed of the following sections.

I. Typical autonomous cleaning device
II. Autonomous Cleaner Behavior Mode
III. Debris sensor structure
IV. Signal processing Conclusion I. Autonomous Cleaning Device Throughout the several figures, like reference numerals are referenced to identify corresponding or similar elements, and FIG. 1 is a schematic plan view of a typical autonomous cleaning robotic device 100. In the device, the debris sensor according to the present invention is incorporated. FIG. 2 is a block diagram of hardware of the robot apparatus 100 of FIG.

  Hardware and action mode (coverage action or movement pattern for clean operation; retreat action to temporary movement pattern; robot apparatus 100 commercialized under the ROOMBA registered trademark by iRobot Corporation of Burlington, MA; An example of safety behavior for emergency situations is described below for a more complete understanding of how the debris sensing system of the present invention is employed. However, the present invention can also be employed in non-autonomous cleaning devices, examples of which are described below in connection with FIG.

In the following description, the terms “forward” and “front” are used to indicate the main direction (forward) of movement of the robotic device (see the arrow indicated by the reference character “FM” in FIG. 1). The front / rear axis FA _x of the robot apparatus 100 coincides with the central diameter of the robot apparatus 100 that divides the robot apparatus 100 into substantially symmetrical right and left halves, each defined as a main side and a non-main side. .

  An example of such a robotic device 100 has a generally disk-shaped storage structure and includes a chassis 102 and an external shell 104 secured to the chassis 102 to define a structural envelope of minimum height. (To facilitate movement under furniture). The hardware comprising the robotic device 100 is typically a power system, an activation power system (also referred to herein as a “drive system”), a sensor system, a control module, a side brush assembly, or a self-adjusting cleaning head system, respectively. They can be classified as basic elements, all of which are combined with the storage structure. In addition to the hardware so classified, the robotic device 100 further includes a front bumper 106 having a generally arcuate configuration and a front wheel assembly 108.

  The front bumper 106 (shown as a single component; or a two-part component) is movably integrated with the chassis 102 (by a displaceable support member pair) and extends outwardly therefrom. ing. Whenever the robotic device 100 collides with an obstacle (eg, wall, furniture) during its movement, the bumper 106 is displaced (compressed) toward the chassis 102 and, once contact with the obstacle is finished, the original stretched. Return to posture.

  The front wheel assembly 108 is mounted in a biased combination on the chassis 102 so that the front wheel assembly 108 is in a retracted position during the cleaning operation (due to the weight of the robotic device 100) and is being cleaned. Rotate freely on the surface. When the front wheel assembly 108 encounters a steep slope (eg, a downstairs, a mezzanine floor) during operation, the front wheel assembly 108 is tilted to an extended position.

  The power system hardware that provides energy to power the electrically powered hardware of the robotic device 100 includes a rechargeable battery pack 110 (and associated conductors) integrated in combination with the chassis 102.

  As shown in FIG. 1, the activation power system provides a means to propel the robotic device 100 during movement of the robotic device 100 and to actuate cleaning mechanisms such as, for example, side brush assemblies and self-adjusting cleaning heads. The activation power system includes left and right main drive wheel assemblies 112L and 112R, their associated independent electric motors 114L and 114R, and electric motors 116 and 118, each for operating a side brush assembly and a self-adjusting cleaning head. I have.

  Electric motors 114L and 114R are mechanically coupled to main drive wheel assemblies 112L and 112R, respectively, as shown in FIG. 1 and in response to execution of a behavior mode or as described in detail below. In response to a debris signal generated by the right debris sensors 125L and 125R, it is independently activated by a control signal generated by the control module.

  Due to the independent operation of the electric motors 114L and 114R, the main drive wheel assemblies 112L and 112R are (1) rotated at the same speed in the same direction to propel the robotic device 100 forward or backward on a straight line, (2) In order to realize various right and / or left rotation patterns (in a range from sharp rotation to shallow rotation) of the robot apparatus 100, the robot apparatus 100 is differentially rotated (one wheel assembly is not rotated). And (3) rotating the robot apparatus 100 in its place, that is, rotated at the same speed in the opposite direction for “rotation on 10 cents of silver coins”. Provide a wide repertoire.

  As shown in FIG. 1, the sensor system includes a variety of different sensor units that operate to generate signals that control the behavior mode operation of the robotic device 100. The robot apparatus 100 includes an obstacle detection unit 120, a cliff detection unit 122, a wheel drop sensor 124, an obstacle tracking unit 126, a virtual wall omnidirectional detector 128, a stall sensor unit 130, a main wheel encoder unit 132, and In accordance with the present invention, left and right debris sensors 125L and 125R are described in detail below.

In the illustrated embodiment, the obstacle (“collision”) detection unit 120 may be an IR break beam sensor mounted in combination with the displaceable support member pair of the front bumper 106. These obstacle detection units 120 are one or more indicating the relative displacement between one or more pairs of support members whenever the robotic device 100 collides with an obstacle such that the front bumper 106 is compressed. Operate to generate more than one signal. These signals, prior to the robotic device 100 - with respect to the rear axle FA _X, in order to determine an approximate point contact with the obstacle (and to be performed behavioral mode) are processed by the control module.

  The cliff detection unit 122 is mounted in combination with the front bumper 106. Each precipice detection unit 122 is configured to operate in an IR emitter / detector configured to establish a focus such that radiation emitted downward from the emitter is reflected from the surface being traversed by the detector and being detected. Provide a pair. If the reflected radiation is not detected by the detector, i.e. a steep slope is encountered, the cliff detection unit 122 sends a signal to the control module (which causes one or more behavior modes to be performed).

  A wheel drop sensor 124, such as a contact switch, is integrated in combination with each of the main drive wheel assemblies 112L and 112R and the front wheel assembly 108, that is, when one of the wheel assemblies is in an extended position, that is, across. Operate to generate a signal whenever it is not in contact with a surface (this causes the control module to execute one or more modes of action).

  The obstacle tracking unit 126 in the above-described embodiment is an IR emitter / detector pair mounted on the “main” side (the right side in FIG. 1) of the robot apparatus 100. The emitter / detector pair is similar in construction to the precipice detection unit 122 but is arranged so that the emitter can emit radiation laterally from the main side of the robotic device 100. Unit 126 operates to send a signal to the control module whenever an obstacle is detected as a result of radiation reflected from the obstacle and detected by the detector. In response to this signal, the control module causes one or more behavior modes to be executed.

  The virtual wall detection system used in conjunction with the above-described embodiment of the robotic device 100 is a stand-alone type that transmits an omnidirectional detector 128 mounted on top of the outer shell 104 and an axially directed confinement beam. A transmission unit (not shown) is provided. The stand-alone transmission unit is a specially designed device in which the emitted confinement beam impedes access to the defined work area, thereby cleaning the robot apparatus 100 with a defined work area (eg, the robot apparatus 100). It is limited to the operation of the entrance). Upon detection of the confined beam, the omni-directional detector 128 transmits a signal to the control module (which causes one or more modes of action to be performed to generate the robotic device 100 by the stand-alone transmission unit. Move away from the confined beam).

  The stall sensor unit 130 is integrated in combination with each electric motor 114L, 114R, 116, and 118, and when a change in current is detected in the associated electric motor (this indicates a fault condition in the corresponding driven hardware) ) And operate to send a signal to the control module. The control module operates to execute one or more behavior modes in response to such a signal.

  A main wheel encoder unit 132 (see FIG. 2) is integrated in combination with each main drive wheel assembly 112L and 112R and operates to detect rotation of the corresponding wheel and correspondingly send a signal to the control module. (The rotation of the wheel can be used to estimate the distance traveled by the robotic device 100).

Control Module Referring now to FIG. 2, the control module includes a sensor of the robotic device 100, a microprocessing unit 135 that includes I / O ports connected to controllable hardware, and a microcontroller (Motorola MC9512E128CPV 16-bit controller). And ROM and RAM memory. The I / O port serves as an interface between the microcontroller and sensor unit (including left and right debris sensors 125, described in more detail below) and controllable hardware, and the signals generated by the sensor unit are transmitted to the microcontroller. And a control (command) signal generated by the microcontroller is transferred to controllable hardware to execute a special behavior mode.

  The microcontroller processes the sensor signal, executes a special behavior mode based on such a processing signal, and controls the controllable hardware based on the behavior mode executed for the robot apparatus 100. Operates to execute a set of instructions for generating (instruction) signals. The clean coverage and control program for the robotic device 100 is stored in the ROM of the microprocessing unit 135, which includes the behavioral mode, sensor processing algorithm, control signal generation algorithm, and control of the robotic device 100 in which behavioral mode or A priority algorithm that determines whether to give to multiple behavior modes is included. The RAM of the microprocessing unit 135 is used to store an activity state of the robot apparatus 100, and an ID of an action mode (a plurality of action modes) in which the robot apparatus 100 is currently operated and a hardware command related thereto are stored. include.

  Referring again to FIG. 1, there is shown a side brush assembly 140 configured and operated to collect particles external to the periphery of the containment structure and direct the particles to a self-adjusting cleaning head system. The side brush assembly 140 provides the robotic device 100 with a function to clean the surface proximate to the bottom plate when the robotic device 100 is operated in the “obstacle tracking” behavior mode. As shown in FIG. 1, the side brush assembly 140 is preferably mounted in combination with the chassis 102 in the front quadrant on the main side of the robot apparatus 100.

  The described self-adjusting cleaning head system 145 for the robotic device 100 comprises a two-stage brush assembly and a vacuum assembly, each powered independently by an electric motor (reference numeral 118 in FIG. Identifies two independent electric motors, one for the brush assembly and one for the vacuum assembly). The cleaning function of the robotic device 100 is commonly characterized by the width of the cleaning head system 145 (see reference character W in FIG. 1).

  With reference now to FIG. 3, in one embodiment of the robotic device 100, the cleaning brush assembly comprises an asymmetric, anti-rotating flapper and main brush elements 92 and 94, respectively, which are in front of the vacuum assembly inlet 84. And operates to direct particulate debris 127 toward the removal dust cartridge 86. As shown in FIG. 3, the robotic device 100 can also include left and right debris sensor elements 125PS, which can be piezoelectric sensor elements, as described in detail below. Piezoelectric debris sensor element 125PS can be located in the cleaning path of robotic device 100, for example, particles 127 mounted on the roof of the cleaning head, thereby being swept up by brush elements and / or attracted by vacuum. In the event of a collision, the debris sensor element 125PS generates an electrical pulse indicating that the debris has collided and, thus, is present in the area where the autonomous cleaning device is operating.

  More specifically, in the arrangement shown in FIG. 3, the sensor element 125PS is substantially located on the axis AX along which the main and flapper brushes 94 and 92 meet so that the particles are at maximum force with the debris sensor element. Collide with 125PS.

  As shown in FIG. 1, and as will be described in more detail below, the robotic cleaning device can be fitted with left and right piezoelectric debris sensors that signal the robotic device to point in the “dirty” area. Generate separate left and right debris signals that can be processed for transmission.

Along with signal processing and selection of a behavior mode based on the debris signal generated by the piezoelectric debris sensor, the operation of the piezoelectric debris sensor is discussed below in accordance with a brief description of the general aspects of the behavior mode of the cleaning device.
II Behavioral Mode The robotic device 100 can employ various behavioral modes to effectively clean the defined work area, where the behavioral mode is a layer of the control system and can operate in parallel. The microprocessing unit 135 performs a prioritized arbitration scheme to identify and execute one or more primary operating modes for any given scenario based on input from the sensor system. To work.

  The behavior mode of the robot device 100 described above is characterized by (1) coverage behavior mode, (2) retreat behavior mode, and (3) safe behavior mode. The coverage behavior mode is mainly designed to allow the robot apparatus 100 to perform its clean operation in an efficient and effective manner, and the evacuation and safety behavior modes are the signals from the sensor system, Normal operation of the device 100 is a priority action mode that is executed when, for example, an obstacle is encountered and inhibited, or a steep slope is detected and is likely to be inhibited, for example.

  Exemplary and exemplary coverage behavior (clean) modes of the robotic device 100 include (1) “spot coverage” pattern, (2) “obstacle tracking (or“ edge clean ”) coverage” pattern, and (3 ) The “room coverage” pattern is included. The “spot coverage” pattern causes the robot apparatus 100 to clean a limited area within the defined work area, for example, an area with a high traffic volume. In a preferred embodiment, the “spot coverage” pattern is performed by a helical algorithm (although other types of self-occluding region algorithms can be used, for example, polygons). A helical algorithm that causes the robotic device 100 to move outwardly in a spiral (preferred) or inwardly is a control signal from the microprocessing unit 135 to the main wheel assemblies 112L and 112R. The rotation radius / multiple radii are varied as a function of time (which increases / decreases the helical movement pattern of the robotic device 100).

  The robotic device 100 may, for example, one or two of the obstacle detection units 120 for a predetermined or random time, for a predetermined or random distance (eg, a maximum helical distance), and / or until a special event occurs. It operates in a “spot coverage” pattern until the above is activated (collectively, transitional state). Once the transitional state occurs, the robotic device 100 may be operated in different modes of operation, such as a “straight line” behavior mode (in the preferred embodiment of the robotic device 100, the “straight line” behavior mode is a low priority, Different behavior modes, such as a default behavior that propels an approximate straight line at a preset speed of approximately 0.306 m / s) or a “bound” behavior mode combined with a “straight” behavior mode. Run or transition to that mode.

  If the transitional state is a result of the robot apparatus 100 encountering an obstacle, the robot apparatus 100 can take another action instead of transitioning to a different action mode. Under the control of the spiral algorithm, the robot apparatus 100 can temporarily execute an action mode to avoid or evacuate an obstacle and resume operation (that is, continue the spiral movement in the same direction). To do). Alternatively, the robot apparatus 100 may temporarily execute an action mode to avoid or retreat from an obstacle and resume operation by controlling the helical algorithm (but in the opposite direction, that is, Reflection helix movement).

  Due to the “obstacle tracking coverage” pattern, the robotic device 100 allows the surrounding of the defined work area, eg, a room partitioned by walls, and / or obstacles (eg, furniture) in the defined work area. Clean the surroundings. Preferably, the robot apparatus 100 of FIG. 1 uses the obstacle tracking unit 126 (see FIG. 1) to continuously maintain its posture with respect to obstacles, such as walls, furniture, and the like. It is possible to move to the vicinity of the obstacle and clean the surroundings at the same time. Different embodiments of the obstacle tracking unit 126 may be used to implement the “obstacle tracking” behavior pattern.

  In the first embodiment, the obstacle tracking unit 126 is operated to detect the presence or absence of an obstacle. In another embodiment, the obstacle tracking unit 126 is activated to detect an obstacle and maintain a predetermined distance between the obstacle and the robotic device 100. In the first embodiment, the microprocessing unit 135 performs a small clockwise or counterclockwise rotation in response to a signal from the obstacle tracking unit 126 to maintain its posture with respect to the obstacle. To work. The robot apparatus 100 performs a small clockwise rotation when the robot apparatus 100 transitions from obstacle detection to non-detection (from reflection to non-reflection), or the robot apparatus 100 performs detection to non-detection (non-detection). When transitioning from reflection to reflection, perform a small counter-clockwise rotation. A similar rotating action is executed by the robot apparatus 100 to maintain a predetermined distance from the obstacle.

  The robotic device 100 may, for example, one or two of the obstacle detection units 120 for a predetermined or random time, for a predetermined or random distance (eg, maximum or minimum distance), and / or until a special event occurs. It operates in the “obstacle tracking” behavior mode until one or more are activated a predetermined number of times (collectively, transitional states). In some embodiments, the microprocessing unit 135 is an obstacle detection unit in an “obstacle tracking” behavior mode that performs a minimum angle counterclockwise rotation to align the robotic device 100 with respect to the obstacle. When 120 is activated, the robot apparatus is caused to execute the “alignment” behavior mode.

  The “room coverage” pattern can be used by the robotic device 100 to clean any defined work area bounded by walls, stairs, obstacles, or other barriers (eg, virtual wall units). A preferred embodiment of the “room coverage” pattern includes a “random bound” behavior mode combined with a “straight” behavior mode. Initially, the robotic device 100 encounters an obstacle under the control of a “linear” behavior mode, ie, a linear algorithm (primary drive wheel assemblies 112L and 112R operating in the same direction at the same rotational speed). Run until. When one or more obstacle detection units 120 are activated, the microprocessing unit 135 calculates an acceptable range in the new direction based on the activated obstacle tracking unit (s) 126. Operate. The microprocessing unit 135 selects a new direction from within an acceptable range and performs a clockwise or counterclockwise rotation to achieve the new direction with minimal movement. In some embodiments, the new direction of rotation continues to move forward, increasing the cleaning performance of the robotic device 100. The new direction may be selected randomly from an acceptable range of directions or based on some statistical selection scheme, eg, a Gaussian distribution. In other embodiments of the “room coverage” behavior mode, the microprocessing unit 135 can be programmed to change direction randomly or a predetermined number of times without input from the sensor system.

  The robotic device 100 may, for example, activate the obstacle detection unit 120 a predetermined number of times until a predetermined or random time, a predetermined or random distance (eg, maximum or minimum distance), and / or a special event occurs. It is operated in “room coverage” mode until it is done (collectively in a transitional state).

  As an example, the robotic device 100 can include four retreat behavior modes: a “rotation” behavior mode, an “edge” behavior mode, a “wheel drop” behavior mode, and a “slow” behavior mode. Those skilled in the art will recognize that other behavior modes can be used by the robotic device 100. One or more of these modes of behavior may be, for example, in one of the side brush assembly 140 or one of the two-stage brush assembly electric motors 116 and 118, raising the current above the low or high stall threshold, for a predetermined time. This is executed in response to the detection of the compressed posture of the front bumper 106 and the wheel fall.

  In the “rotate” behavior mode, the robotic device 100 starts in a random direction at a higher speed (eg, twice the nominal rotational speed) and decreases to a lower speed (half the nominal rotational speed), ie, each Rotate in place with small panic rotation and large panic rotation. Preferably, the small panic rotation is in the range of 45 ° to 90 °, and the large panic rotation is preferably in the range of 90 ° to 270 °. In the “rotation” behavior mode, under the obstacle in the room, the robotic device 100 becomes unable to move on the obstacle in the room, for example, on the high part of the carpet, on the ramp ramp. For example, avoid getting stuck under the sofa or being trapped in a narrow area.

  In the “edge” behavior mode, none of the obstacle detection units 120 is activated, until the robot device rotates by a predetermined angle, for example, 60 °, or from the start of the “edge” behavior mode. Obstacle edges are tracked until rotated by an angle, for example, 170 °. The “edge” behavior mode allows the robotic device 100 to pass through the smallest possible hole to retreat from a narrow area.

  In the “wheel drop” behavior mode, the microprocessing unit 135 temporarily reverses the direction of the main wheel drive assemblies 112L and 112R and then stops it. When the activated wheel drop sensor 124 stops its activity within a predetermined time, the microprocessing unit 135 re-executes the action mode that was being executed before the wheel drop sensor 124 was activated.

  In response to an event, for example, the activation of the wheel drop sensor 124 or the cliff detection unit 122, a “slow” behavior mode is executed for a predetermined distance to decelerate the robotic device 100 and then gradually to the original operating speed. Return to.

  When a safe state is detected by the sensor subsystem, for example, a brush or wheel stall that causes the corresponding electric motor to temporarily turn off occurs continuously, or the wheel drop sensor 124 or the cliff detection unit 122 When the robot apparatus 100 is activated for a predetermined time, the robot apparatus 100 is generally turned off. In addition, an audible alarm is generated.

The above description of the behavior modes for the robotic device 100 is merely a representation of the types of operational modes that can be executed by the robotic device 100. Those skilled in the art will recognize that the behavior modes described above can be implemented in other combinations and / or situations, and that other behavior modes and movement patterns are also possible.
III Structure and Operation of Debris Sensor As shown in FIGS. 1 to 3, according to the present invention, the autonomous cleaning device (and similarly, the non-autonomous cleaning device shown as an example in FIG. 7) incorporates the debris sensor. Can be improved. In the embodiment illustrated in FIGS. 1-3, the debris sensor subsystem includes left and right piezoelectric debris sensor elements 125L and 125R located near or within the cleaning path of the cleaning device, and debris from the sensor. It comprises electronics elements for processing the signal and transferring it to the microprocessing unit 135 or other controller.

  When employed in an autonomous robotic cleaning device, the debris signal from the debris sensor selects the behavior mode (to be in spot cleaning mode), changes the operating state (such as speed, power, or others) It can be used to steer in the direction of debris (especially, equally spaced left and right debris sensors are used to generate differential signals) or to take other actions.

  The debris sensor according to the present invention can also be incorporated into a non-autonomous cleaning device. For example, when employed in a relatively non-autonomous cleaning device such as the vacuum cleaner 700 belonging to the prior art as shown in FIG. 7, the debris generated by the piezoelectric debris sensor 704PS located in the cleaning or vacuum path of the device. The signal 706 is employed by the control microprocessor 708 in the body of the vacuum cleaner 702 to generate a signal that can be recognized by the user (eg, by turning on the indicator light 710) to increase the power from the power system 703; Or some combination of actions (such as turning on a "high power" lamp and simultaneously increasing power) can be taken.

  The algorithm aspects of the operation of the debris sensor subsystem are summarized in FIG. As shown therein, the method according to the invention detects collisions with debris and left and right debris signals indicating the presence, quantity or volume and direction of debris (802), debris signal values. To select an operating mode or movement pattern (such as “spot coverage”) based on 804, to select a movement direction based on the differential left / right debris signal (eg, to the more debris side) (806), generating a user-recognizable signal indicating the presence of debris or other characteristics (eg, lighting a user-recognizable LED) (808), or operating state, For example, changing or controlling power (810) can be included.

  A further practice of the present invention utilizes the movement of the autonomous cleaning device on the floor or other surface to process the debris signal in conjunction with information about the movement of the cleaning device to calculate the debris gradient (FIG. 812 in 8). The debris gradient shows the change in the collision count with debris as the autonomous cleaning device moves along a surface. By examining the sign of the slope (positive or negative in relation to increasing or decreasing debris), the autonomous cleaning device controller continuously adjusts the device path or movement pattern to most effectively locate the debris. It can be cleaned (812).

Piezoelectric Sensor As noted above, piezoelectric transducer elements can be used in the debris sensor subsystem of the present invention. Piezoelectric sensors have a relatively high immunity to adhesion that responds instantaneously to debris impact and degrades the performance of optical debris sensors typical of the prior art.

  An example of a piezoelectric debris sensor element 125PS is shown in FIG. Referring to FIG. 4, the piezoelectric debris sensor element 125PS has one or more electrodes, each having a brass disk 402 made of piezoelectric material having a thickness of 0.20 mm and a diameter of 20 mm, connected to the top side thereof. (Total thickness 0.51 mm), which in turn are mounted on an elastomeric pad 404, a plastic dust sensor cap 406, a debris sensor PC board including associated electronics elements 408, and a grounded metal shield 410. And is supported by mounting screws (or bolts or the like) 412 and elastomeric grommet 414. The elastomeric grommet 414 attenuates the vibrations to some extent and provides insulation between the piezoelectric debris sensor element 125PS and the cleaning device.

  In the example shown in FIG. 4, a rigid piezoelectric disk of the type typically used for inexpensive sounders can be used. However, a flexible piezoelectric film can also be advantageously employed. Since the film can be manufactured in any shape, its use offers the possibility of sensitivity to debris over the entire cleaning width of the cleaning device, rather than sensitivity at the selected location where the disk is located, for example. However, conversely, at present, films are more expensive and tend to deteriorate over time. Conversely, brass discs have proven to be very strong.

  The exemplary implementation shown in FIG. 4 is substantially optimized for use on a mechanically noisy platform, such as an autonomous vacuum cleaner as shown in FIG. In such devices, vibration attenuation or sensor isolation is very effective. However, in an application example including a non-autonomous cleaning device such as a box-type vacuum cleaner as shown in FIG. 7, the damping mode of the mounting system of FIG. 4 is unnecessary. In non-autonomous cleaning devices, an alternative mounting system may include heat to secure the piezoelectric element directly to its housing. In any case, an important consideration for achieving improved performance is to reduce the area required to maintain the piezoelectric element in place with clamps, bolts, or other methods. The smaller the space occupied by this clamped “dead zone”, the more sensitive the piezoelectric element.

  In operation, debris thrown up by a cleaning brush assembly (eg, brush 94 of FIG. 3) or flowing through a cleaning path (eg, vacuum component 104 of FIG. 3) within the cleaning device is piezoelectric debris sensor element 125PS. May collide with the bottom of the wall, that is, all the brass parts. As shown in FIG. 3, in the robot apparatus 100, the piezoelectric debris sensor element 125PS is substantially located on an axis AX along which the main brush 94 and the flapper brush 92 meet, whereby the fine particles 127 are thrown up. And collides with the piezoelectric debris sensor element 125PS with the maximum force.

  As is well known, piezoelectric sensors convert mechanical energy (for example, kinetic energy from debris collisions and brass disk vibrations) into electrical energy (in this case, an electrical pulse is generated each time the debris collides. Is processed and transmitted to the system controller (eg, microprocessing unit 135 of FIGS. 1 and 2 or 708 of FIG. 8) to control or change the change of the operating mode according to the present invention. It is a pulse. Piezoelectric elements are typically designed to be used as audio transducers, for example, to generate beep tones. When AC voltages are applied, they mechanically vibrate in accord with the AC waveform to produce an audible output. Conversely, when they mechanically vibrate, they produce an AC voltage output. In this way, they are employed in the present invention. In particular, when the object first strikes the brass surface of the sensor, the disk is bent inward, thereby generating a voltage pulse.

Filtering However, since the piezoelectric debris sensor element 125PS is in direct or indirect contact with the body through the cleaning device chassis, or its mounting system (see FIGS. 3 and 4), when the cleaning device is functioning, the motor, Susceptible to mechanical vibrations normally generated by brushes, fans, and other moving parts. This mechanical vibration causes the sensor to output an unwanted noise signal that is larger in amplitude than the signal generated by small low mass debris that impacts the sensor (such as crushed black pepper). The net result is that the sensor generates a composite signal composed of low frequency noise components (up to about 16 kHz) and higher frequency, preferably smaller amplitude, debris impact components (greater than 30 kHz, up to several hundred kHz). Output. Thus, it is advantageous to provide a means for filtering unimportant signals.

Thus, as described below, electronic filters are used to significantly attenuate low frequency signal components and improve signal to noise performance. Examples of the architecture and circuitry of such filtering and signal processing elements are described in connection with FIGS.
IV Signal Processing FIG. 5 is a schematic diagram of the signal processing elements of the debris sensor subsystem in one practice of the invention.

  As mentioned above, one purpose of the debris sensor is to allow the autonomous cleaning device to sense when it picks up or encounters a debris location. This information is used as an input to make changes in the cleaning behavior or to allow the device to enter a selected operation or behavior mode, such as the spot cleaning mode described above when encountering debris, for example. it can. In the non-autonomous cleaning device shown in FIG. 7, the debris signal 706 from the debris sensor 704PS turns on an indicator lamp 710 that can be recognized by the user (for example, signals to the user that debris is encountered), Used to increase the power output from the power unit 703 to the cleaning system, or to cause other operational changes or combinations of changes (eg, turn on a “high power” light that the user can recognize and increase the power at the same time) it can.

  In addition, as described above, the two debris sensor circuit modules (ie, the left and right channels such as 125L and 125R in FIG. 1) are capable of debris volume picked up by the autonomous cleaning device on the right and left sides of the cleaning head assembly. Can be used to be able to perceive the difference between. For example, if the autonomous cleaning device encounters a garbage location ahead of its left side, the left debris sensor indicates a debris collision and the right sensor does not indicate a debris collision (or a small amount if shown). This differential output is driven by a microprocessor controller (such as the microprocessing unit 135 in FIGS. 1 and 2) of the autonomous cleaning device that steers the device in the direction of debris (for example, the left debris sensor has a higher signal than the right debris sensor). It can be used to steer to the left when generating values), or to select a vector of debris directions, or to select a movement or behavior pattern such as spot coverage or others.

  Thus, FIG. 5 illustrates one channel (eg, left channel) of the debris sensor subsystem that can include both left and right channels. The right channel is substantially identical and therefore its structure and operation will be understood from the discussion below.

  As shown in FIG. 5, the left channel includes a sensor element (piezoelectric disk) 402, an acoustic vibration filter / RFI filter module 502, a signal amplifier 504, a reference level generator 506, an attenuator 508, an attenuator 508, and a reference level generator. It comprises a comparator 510 that compares the outputs of 506 and a pulse stretcher 512. The output from the pulse stretcher 512 is a logic level output signal to the microprocessing unit 135 shown in FIG. 2, ie, a system controller such as a controller suitable for use in selecting operational behavior.

  The acoustic vibration filter / RFI filter block 502 provides significant damping (in one embodiment, greater than -45 db volts), blocks most of the low frequency, low rate of mechanical signals, and high frequency. It can be designed to pass a high rate of change debris collision signal. However, even if these high-frequency signals pass through the filter, they are attenuated and thus need to be amplified by the signal amplifier 504.

  In addition to the amplification of the desired high frequency debris collision signal, a very small residual mechanical noise signal passing through the filter is also generated by the motor itself and radiated in the atmosphere by the electrical noise generated by the amplifier itself. Or amplified with any radio frequency interference (RFI) component picked up by the sensor and its leads. The high frequency response of the signal amplifier is designed to minimize the amplification of very high frequency RFI. This constant background noise signal having a frequency component much lower than the desired debris collision signal is provided to the reference level generator 506. The purpose of the reference level generator 506 is to generate an instantaneous peak value of the noise signal, or a reference signal following the envelope. The signal of interest in FIG. 5, ie the signal resulting from the debris colliding with the sensor, is also supplied to this reference level generator 506. Thus, the reference level generator 506 is designed to track the instantaneous peak values of these signals without responding so quickly to high frequency, high rate of change debris collision signals. The resulting reference signal is used for comparison as described below.

  Referring again to FIG. 5, it can be seen that the signal from the signal amplifier 504 is also supplied to the attenuator 508. Since this is the same signal supplied to the reference level generator 506, it is a decoded signal that includes both the high frequency signal of interest (ie, when debris collides with the sensor) and low frequency noise. The attenuator 508 reduces the amplitude of this signal, typically less than the amplitude of the signal from the reference level generator 506 when the debris is not hitting the sensor element.

  Comparator 510 compares the instantaneous voltage amplitude value of the signal from attenuator 508 with the signal from reference level generator 506. Normally, when the cleaning device is in operation and debris has not collided with the sensor element, the instantaneous voltage coming out of the reference level generator 506 is higher than the voltage coming out of the attenuator 508. This causes comparator 510 to output a high logic level signal (logic 1), which is inverted by pulse stretcher 512 to produce a low logic level (logic 0).

  However, when the debris collides with the sensor, the voltage from the attenuator 508 exceeds the voltage from the reference level generator 506 (this circuit cannot track the signal component with a high rate of change at a high frequency from the signal amplifier 504. Therefore, the signal generated by the debris collision is higher in voltage amplitude than the constant background mechanical noise signal that is more severely attenuated by the acoustic vibration filter 502. This causes comparator 510 to temporarily change the state to logic level 0. The pulse stretcher 512 extends this very short (typically less than 10 microseconds) event to a constant 1 millisecond (+0.3 mS, −0 mS) event, and the system controller (eg, the microprocessing unit 135 of FIG. 2). ) Provides sufficient time for signal sampling.

  When the system controller "sees" this 1 millisecond logic 0 pulse, it interprets the event as a debris collision.

  Referring now to the RFI filter portion of the acoustic vibration filter / RFI filter 502, this filter functions to attenuate electrical noise (RFI) radiated at very high frequencies generated by the motor and motor drive circuitry. To do.

  In summary, the example circuit connected to the sensor element uses both amplitude and frequency information to indicate a debris collision (which indicates a cleaning device picking up the debris), which was also picked up by the sensor element. Distinguish between normal background mechanical noise and radiated radio frequency electrical noise generated by the motor and motor drive circuitry. Although not desirable, normal constant background noise is used to establish a dynamic reference that avoids false debris collision symptoms while maintaining a good signal-to-noise ratio.

  In practice, the mechanical mounting system for the sensor element (see FIG. 4) is also designed to minimize mechanical acoustic noise vibration coupling that affects the sensor element.

Signal Processing Circuit (consisting of FIGS. 6A to 6C) FIG. 6 is a detailed schematic diagram of a typical debris sensor circuit. Those skilled in the art will appreciate that in other embodiments, signal processing may be performed partially or entirely within the software of the microprocessing unit 135. Referring to FIG. 6, the illustrated example of a suitable signal processing circuit includes the following elements and operates in accordance with the following description.

  A composite signal based on ground from a piezoelectric sensor disk (see piezoelectric disk 402 in FIG. 4) is supplied to the capacitor C1 and attenuates low-frequency acousto-mechanical vibrations that are directed to the sensor via the mounting system. Input to a 5-pole, high-pass passive R-C filter. The filter has a -3 dB corner frequency that rolls off at 2-100 kHz, -100 dB / Decade (decrease rate for a 10-fold change in frequency). The output of this filter is fed to a single pole, low pass passive RC filter designed to attenuate any very high frequency RFI. This filter has a −3 dB corner frequency that rolls off at −20 dB / Decade of 1.06 MHz. The output of this filter is diode clamped by D1 and D2, protecting the sensor element from U1 high voltage transitions in the event of persistent collisions that generate voltage pulses larger than the amplifier supply rail. The DC bias required for the signal supply operation for the amplifier chain and the subsequent comparator circuit is generated by R5 and R6. These two resistance values are chosen so that their Thevenin impedance works with C5 to correctly maintain the fifth pole frequency response of the filter.

  U1A, U1B, and their associated components theoretically form a two-stage, ac-coupled, non-inverting amplifier with an AC gain of 441. C9 and C10 function to minimize gain at low frequencies, while C7 and C8 function to roll off gain at RFI frequencies. The net theoretical frequency response from the filter input to the amplifier output is 32.5 kHz, -100 dB / Decade with a -3 dB single pole high pass response, 100 kHz, -32 dB / Decade, and 5.4 MHz, -100 dB / In Decade, it is a two-pole low-pass response having a break frequency and together forms a band-pass filter.

  The output from the amplifier is split and one output is input to R14 and the other is input to the non-inverting input of U1C. The signal input to R14 is attenuated by the R14-R15 voltage divider and supplied to the inverting input of the comparator U2A. The other signal branch from the output of U1B is fed to the non-inverting input of amplifier U1C. U1C along with U1D and the components in between (as shown in FIG. 2) form a half-wave, positive peak detector. The attack and decay times are set by R13 and R12, respectively. The output from this circuit is supplied to the non-inverting input of U2A via R16. R16, along with R19, provides hysteresis to improve switching time and noise immunity. U2A functions to compare the instantaneous value between the output of the peak detector and the R14-R15 attenuator.

  Normally, when the debris does not strike the sensor, the peak detector output is greater in amplitude than the attenuator network output. When the debris strikes the sensor, a high frequency pulse is generated that has an amplitude that is larger than the low frequency mechanical noise signal component and exits the front end high pass filter and enters U1A. This signal is greater in amplitude than it exits the peak detector, even after leaving the R14-R15 attenuator network, which means that the peak detectors of R13, C11, and R12 This is because high-speed pulses cannot be tracked due to component values in the network. The comparator changes state from high to low as long as the amplitude of the debris collision pulse is greater than the dynamic, noise-generated, reference level signal exiting the peak detector. Since this comparator output pulse is too short for the system controller to see, a pulse stretcher is used.

  The pulse stretcher is a one-shot monostable design with a lockout mechanism, thereby avoiding retriggering until the end of the timeout period. The output from U2A is supplied to the junction of C13 and Q1. C13 couples the signal monostable formed by U2C and its related components. Q1 functions as a lockout by keeping U2A output low until monostable times out. The timeout period is set by the time constant formed by R22, C12, and R18 and the reference level set by the R20-R21 voltage divider. This time can be adjusted for 1 mS, +0.3 mS, and -0.00 mS, as defined by the software requirements used by the controller / processor.

  The power of the debris sensor circuit is provided by U3 and related components. U3 is a low power linear regulator that provides a 5 volt output. Unregulated voltage from a battery mounted on the robot provides power input.

  If necessary, circuit adjustment is set by R14 and R12. With these adjustments, the circuit response can be tuned in a short time.

In this type of generator, power to the debris sensor circuit printed circuit board (PCB) and signals from it are transferred to the main board via a shielded cable. Alternatively, the noise filter can be replaced with the use of a shielded cable, reducing the wiring cost. The cable shield drain wire can be grounded only on the sensor circuit PCB side. This shield does not carry any ground current. Separate conductors in the cable carry the power ground. For noise reduction, the production sensor PCB should have all components on the top side with a solid ground plane on the bottom side. The sensor PCB should be housed in a mounting assembly with a grounded metal shield that covers the top side of the substrate to shield components from radiated noise pickup from the robot motor. The piezoelectric sensor disk can be mounted under the sensor circuit PCB on a suitable mechanical mounting system, as shown in FIG. 4, so that the connection leads are as short as possible for noise immunity. Can be maintained.
V. CONCLUSION The present invention is capable of instantaneously sensing and responding to debris collisions without being degraded by debris deposition, and thus responding directly to debris on the floor or other surface to be cleaned. For low fluctuations, instantaneous power, or other operating conditions of the cleaning device, a low sensitivity debris sensor is provided.

  When employed as described herein, according to the present invention, an autonomous cleaning device controls its operation or, for example, directs the device to a more “dirty” location based on a signal generated by a debris sensor. Can be selected from a plurality of operating modes, movement patterns, or behaviors that respond to detected debris.

  The debris sensor can also be employed in a non-autonomous cleaning device and can control, select or change the mode of operation of either the autonomous or non-autonomous cleaning device.

  Furthermore, the disclosed signal processing architecture and circuit is particularly effective in combination with a piezoelectric debris sensor and provides a high signal to noise ratio.

  Those skilled in the art will recognize that variations and modifications of the broad scope of the invention are possible within the scope of the invention. Debris sensors can also be employed for purposes and devices other than those described herein. Accordingly, what has been described above is provided by way of example only and the scope of the present invention is limited only by the accompanying claims.

Claims (1)

A cleaning device,
A clean path for transporting debris;
A brush assembly operable to sweep debris toward the cleaning path;
A debris sensor located in a location where the debris swept by the brush assembly collides in proximity to the cleaning path and generates a debris signal indicative of the collision in response to the collision with the debris;
A processor for changing an operation mode of the cleaning device in response to the debris signal;
With
The debris sensor
A filter that attenuates frequency components corresponding to background mechanical noise and electrical noise caused by movement of the cleaning device;
And a determination unit that determines occurrence of collision of the debris from amplitude information in an output from the filter.

Images