MXPA96004777A - nd. AUTOMOTIVE OCCUPANT SENSORY SYSTEM AND METHOD OF OPERATION THROUGH FUSION OF SENSO - Google Patents

Abstract

A system to detect the classification of the presence, position and type of an occupant in a passenger seat of a vehicle, as well as to detect the presence of a child seat that he sees backwards in it, to be used in the control of a system activating control of a related airbag, to allow, disable or control the inflation rate or amount of inflation of an airbag. The detector system employs fusion of sensors, a process to combine the information provided by two or more sensors (24, 26) each of which "observes" the world in a unique sense. In a preferred embodiment, the infrared sensor inputs (78) and the ultrasonic sensor inputs (79) are combined in a microprocessor by a sensor fusion algorithm (80) to produce an occupancy state output signal (85) for the ai bag driver

Description

AUTOMOTIVE OCCUPANT SENSORY SYSTEM AND OPERATING METHOD THROUGH SENSOR FUSION DESCRIPTION CROSS REFERENCE WITH THE RELATED APPLICATION This application is a continuation request in part of the US application SN 08 / 227,531 of the same title, presented by us on April 12, 1994 TECHNICAL FIELD This invention relates to automotive occupancy sensory systems (AOS) and methods of operation by melting sensors to determine the presence and position of an object in a seat, and classifying it by type or nature, in order to provide a state of occupancy or condition signal for use with other automotive control systems, typically in conjunction with an airbag activation or other type of safety restraint system for the protection of passengers in the event of a collision. A main mode is a multiple sensor US / IR occupant detection unit, co-located in the vehicle performer, which provides an occupancy status signal to an air bag controller. The OSA determines by fusion of crossed correlated sensory characteristics the presence, absence, orientation and classification of an animal or human occupant, child seat (that looks back or towards the front), occupant out of position, or other types of occupation, in order to provide an occupancy status signal to the airbag controller, which indicates how appropriate it is to deploy (or not) the airbag, thereby increasing the reliability and safety of an airbag activation system . PREVIOUS TECHNIQUE Virtually all vehicles, cars, cargo vans, and modern trucks have airbag deployment systems. A fraction of increase in the currently available air bag deployment systems includes a passenger side air bag as well as an air bag on the driver's side. However, the passenger-side air bag deployment system presents problems regarding the criteria for its deployment. That is, it is not simply an article that always displays a passenger airbag, as it can cause damage to the occupants through its deployment in certain situations. For example, the airbag should only be deployed if an occupant is actually in the passenger seat, and not when the seat is empty. An even more important problem is the danger of deploying an airbag on the passenger side when it has a rear-facing child seat (RFCS). The deployment of an air bag against the back of an RFCS occupied by a child can cause serious harm to the child by throwing the child toward the back of the car seat, thereby overriding the safety advantages of both the airbag and the child. of the RFCS during a collision. In accordance with the above, it is very important to provide a means to determine when the passenger seat is occupied and when it is not occupied. It is even more important to determine the classification of the "object" in the seat, including when it is occupied by a child in an RFCS so that such information can be used to prevent the deployment of the airbag for an occupancy status of a child. RFCS. Any means to determine the nature and condition of an occupant, including the presence and orientation of a child seat, must be highly reliable in order to signal the selective deployment of the airbag when the seat is occupied by a passenger and avoid the deployment of the airbag when the seat is occupied by an RFCS. This is made up of the fact that there are over thirty-five different infant seats available. The seats are adjustable and the interior configuration and installation of lateral bending of each vehicle is different. In this way, it is not an easy task to provide a sensory system, meaningful sensory units and methods of operation and signal processing, to reliably detect the change of state from an empty seat to a occupied seat and determine the nature (classify), position (location) ) and / or orientation of an object or passenger in the vehicle. As an example, if a thermal sensor is used, its reliability can be reduced by the thermal conditions inside the vehicle, which can change dramatically with the seasons, the weather, the interior configuration of the vehicle, the rapidly changing external shadow, the clothing and / or dimension of the passenger, choice of the driver of the indoor climate, smoking activity, hot food (eg, pizza) on the seat, etc. In this way, a thermal sensor that acts can only lead to a falsely declared occupant presence, and more importantly, to the failure to detect the presence of an occupant. In addition, there may be cases where the thermal identification of an RFCS mixes so well with the seat upholstery that a thermal sensor does not see it, allowing the airbag system to deploy despite the presence of an RFCS occupied by a boy.

Conversely, if one were to use distance measurements instead, such as through the use of acoustic sensors, such a sensor should be able to distinguish between the presence of an RFCS and the presence of a passenger holding an object, being in a position or making a movement that can result in distance measurements that simulate the presence of an RFCS. There are also other scenarios that require a sensory system to recognize, classify and signal air bag controllers to take appropriate action, such as an FFCS, inanimate objects, a passenger holding an inanimate object, a passenger out of position , and so on. In addition to these basic sensory requirements, the system for determining the presence of a passenger in the passenger seat and the presence or absence of a rear-facing child seat must be cost-effective and must be in a package small enough to avoid its interference with the normal operation of the vehicle. Such systems must be compatible with the aesthetics of the vehicle so as not to affect the possibility of selling the vehicle, particularly as it relates to new passenger cars. In addition, the cost of installing such a system in the vehicle must remain simple in order to maintain the low cost of manufacturing. Preferably, all sensors should be stored in a single unit to facilitate the assembly of the vehicle in the production or modernization of older vehicles. There is no currently available sensory system known to the Requesters, which can reliably distinguish the presence, absence and nature of an object or passenger in the passenger seat. No one can currently distinguish selectively the presence or absence of a child seat that looks back on the passenger seat. Nor is there a sensory system currently available that can count for a wide variety of possible variations in both thermal and distance parameters that are found in the wide range of real occupation circumstances, none that is sufficiently versatile to be able to adapt to the wide range of interior configurations of the vehicle. An example of a system for driving a driver's air bag restriction is shown in White's US Patent and other 5,071,160 (Automotive Systems Laboratory), which employs an ultrasonic acoustic sensor to detect the position of the driver, a sensor ". pyrotechnic "to detect the presence of the driver, and a pressure transducer inside the seat to detect the approximate weight of the driver and an airbag control module to activate the deployment of the airbag. As it is better understood, when an imminent collision is detected by a crash sensor, a control module shows the detected position of the passenger at fixed time intervals to calculate the proportion or movement of the passenger in relation to the various fixed motion structures of the passenger. vehicle. This relative movement rate of the passenger is used to corroborate the acceleration data from the crash sensor and ensure deployment of the airbag when the passenger is at substantial risk of damage. That is, the acceleration of the interior passenger is apparently used to prevent false crash signals from the crash sensor. Crash sensors can trigger the deployment of the airbag during a minor hit in near slow traffic or during parking. This "passenger-accelerating-at-same-time" system addresses the correction of false signals from the crash sensor. The patent describes the desired results but does not detail the process or circuitry to achieve these results beyond the establishment that the airbag control circuit uses error correction methods such as a plurality of each type of sensor (sensor). shocks, pyrotechnic, ultrasonic, acoustic and pressure transducer) for each assigned monitoring task in order to prevent falsification. According to the above, it is said that the control circuit employs redundant sensors for each monitoring task and it is said that the instructions executed by the control module include error correction subroutines known to a person skilled in the art. A dashboard signal lamp may come on when the effectiveness of the airbag is too low, or the probability of passenger damage through the airbag is greater than the damage if it is hit with the steering wheel, dashboard or Cross of the knees, being the last one consistent with the situation of slow blow described above. According to the foregoing, there is a need in the art for a reliable occupant sensory system to be used in conjunction with vehicular air bag deployment systems. There is also a need for a sensory system that can meet the aforementioned requirements for reliability in the detection of the presence, absence and classification of an object, a passenger or RFCS in a wide range of circumstances, regardless of whether the passenger Holds an object and without taking into account the thermal conditions that may be found in the vehicle. Such a sensory system must also be an economic component of the vehicle that does not diminish the aesthetics of the interior of the vehicle or unduly increase the cost of manufacturing or assembling a vehicle. EXHIBITION OF THE INVENTION OBJECTS AND ADVANTAGES It is an object of the present invention to provide a sensory automotive occupancy system to reliably detect the presence or absence of a passenger in the passenger seat and the presence or absence of a child seat facing rearward in the seat. of the passenger and provide an occupancy status signal to the airbag control system to allow either the inhibition or deployment of an airbag on the passenger side during a collision. Another object of the present invention is to provide a vehicular passenger detection system that depends on multiple sensors with cross-correlation of characteristics from different physical phenomena to provide signals that are processed by sensor fusion in order to significantly improve the reliability of detection of a passenger while allowing the use of conventional sensors of relatively low cost. It is another object of the invention to provide a vehicle occupancy detection system adapted to be used with a passenger seat of a vehicle in order to control the deployment of an airbag, and specifically to provide an occupancy status signal allowing the control, including inhibition, of deploying an airbag when a passenger seat is classified as unoccupied, occupied by inanimate objects, the occupant is out of position, or when an RFCS is present in the passenger seat, with object of avoiding unnecessary or insecure deployment that might otherwise cause damage. Another object of the invention is to provide a sensory system for occupancy of a passenger that uses both acoustic and thermal sensors, the signals of which are processed in a fusion algorithm to produce an output signal indicative of status and occupancy classification, whose The signal can be used in an air bag control system to allow the deployment of a passenger-side air bag only when the passenger seat is occupied by a passenger properly placed in the seat and inhibit the deployment of a passenger bag. air in other preselected occupation conditions.

Another object of the invention is to provide an occupancy detection system by multiple sensors, which processes by means of sensor fusion certain preselected characteristics extracted from the signals provided by different sensors that detect different physical parameters and correlate them to increase the reliability of the characteristics of individual sensors of the individual sensors. Another object of the invention is to provide a detection system for occupancy by multiple sensors while maintaining low cost in the manufacture of the vehicle by locating multiple sensors co-located in a single unit to facilitate the task of mounting the sensory system in the vehicle . Another object of the present invention is to provide a occupancy detection system by multiple sensors while maintaining the aesthetics of the vehicle by producing a sensory system of minimal size. Another object of the present invention is to provide a system that can be adjusted to the interior configurations of the vehicle with non-parallel discrimination accuracy by processing the sensor fusion signal to produce state, condition or decision signals that can be used as input in a wide variety of automotive systems, including but not limited to occupant safety, vehicle integrity and safety, condition or position of vehicle operation systems (eg, seat position and load adjustment systems), unusual conditions, control of the interior temperature, unauthorized entry (Passive Thief Deterrence), near-object detection systems, and the like. Still other objects will be evident from a review of the Summary, Drawings, Detailed Description and Claims. BRIEF DESCRIPTION OF THE DRAWINGS The invention is illustrated in the drawings, in which: Figures 1-8 show various illustrative conditions of some of the variety and range of real conditions that must be detected and discriminated (classified) accurately by a fully functional automotive occupant sensory system which, by way of example, focuses on a passenger seat of a vehicle, with: Figure 1 showing the seat being occupied by a passenger; Figure 2 showing the passenger seat unoccupied and detected as "empty"; Figure 3 showing a child in a rear-facing child seat ("RFCS"); Figure 4 showing a passenger holding a grocery bag; Figure 5 showing a child in a child seat facing forward ("FFCS"); Figure 6 showing a dog in the seat; Figure 7 showing a passenger out of position ("OOP"); and figure 8 showing a package of moderate size on the seat; Figure 9 is an enlarged front view of the sensor taken along line 9-9 of Figure 2 having an infrared multi-element sensor and an ultrasound sensor contained within a single unit, and illustrating a lens system Fresnel of multiple elements on a colocalized dual detector infrared sensor; Figure 9B is a longitudinal sectional view of the IR sensor taken along the line 9B-9B of Figure 9A; Figure 9C is a cross-sectional view of the IR sensor taken along the line 9C-9C of Figure 9A; Figure 10 is a view of the passenger seat and sensory unit in relative relation, illustrating the area of the infrared detector of the seat and the rear areas of the seat as detected through the Fresnel lenses; FIG. 1 a is a side view illustrating the infrared detector fields of sight coverage on the passenger seat; Figure 11b is a side view illustrating a typical ultrasound transducer field of sight coverage on the passenger seat; Figure 12 is a schematic diagram of the electronic circuit of a modality of the sensory system of the present invention; Fig. 13 is a functional block diagram of an application-specific integrated microcircuit means ("ASIC") for carrying out the sensor fusion methods of the present invention; Figure 14 is a functional block diagram of the signal processor illustrating the processing steps used in the operation of the currently preferred best mode mode of the sensory system of the present invention; Figures 15a and 15b are block diagrams of characteristic processing showing the processing steps of raw data from the sensors to produce infrared (Figure 15a) and ultrasound characteristic vectors (Figure 15b); Figure 16 is a block diagram of processing of fused features illustrating the process of fusing infrared characteristics and ultrasound features to produce a fused feature vector; Fig. 17 is a block diagram of the detection processing showing the processing of the infrared characteristic vector, the ultrasound characteristic vector, and the fused feature vector in order to produce the state; Figure 18 is a graph illustrating the relationship between a feature vector component and the security levels of various occupancy states by way of example: OOP status, RFCS state, inanimate object state, occupant state, and state empty; Figure 19 is a graph illustrating the progress of security levels for a given state and a vector component of a given feature over time; The fi xure 20 is a graph showing the security level after the merging of two feature vector components; Figure 21 is a graphically illustrated matrix of the relationship between the vector components, and the fused vector components, states and security levels; and Figure 22 is a block diagram of decision processing illustrating the factors considered in a state change decision process; Figure 23 is a diagram of the sensory decision reliability in a case of discrimination between (classification of) a normal occupant and an RFCS; Figure 24a shows an identification sign of a car; Figure 24b shows the physical arrangement of the vehicle that gives the indication of Figure 24a; Figure 25a shows an identification sign of a truck; Figure 25b shows the physical arrangement of the vehicle that gives the indication of Figure 25a; The fiction 26 is a table of the test data of the actual examination of a sensory system of the invention; Figures 27a and 27b are comparative indicia showing the sensitivity of discrimination between an RFCS and the same RFCS covered with two blankets; and Figure 28 is an isometric view of a currently preferred embodiment of the AOS unit having co-located, three US and 2 IR sensors inserted in conical bore holes off the axis, to accommodate their respective views. SUMMARY The present invention is directed to an automotive interior occupant sensory system that employs sensor fusion signal processing that combines the information provided by two or more sensors, each of which "sees" the world in a unique sense. The multi-sensor fusion process of this invention greatly improves performance and reliability in almost the same way that the human ability to distinguish and visually classify objects is greatly enhanced with the addition of sound. Although the invention is described in detail with respect to the detection of the presence (or absence) of a variety of seat occupants for the purpose of sending an occupancy status signal to an airbag deployment control system, allowing so that this enables or disables the airbag system to allow or prevent deployment in preselected situations, the "decision" or status signal produced by the sensory system apparatus and the method of fusion of sensory signals of this invention may be applied to also, or alternatively, check, affect or fire other systems, such as automatic safety belts, seat positioning systems, interior climate controls, lighting, dashboard, or other warning signs or lights, warning audio or status signals (electric bells, registers, or synthesized voices), door locks, load adjustment systems, warning systems, shock condition registration systems, and the like. In a preferred embodiment, the passenger seat occupancy sensor of the automobile of the present invention depends on two detectable properties: one such property is the thermal identification and the associated movement, and the second is the acoustic distance and the associated acoustic movement . By relying on a plurality of two different types of sensors in which a plurality of independent configurations (features) are extracted and cross-correlated, and the fusion of some of these characteristics, the accuracy and reliability of the detection is vastly improved in comparison to a single sensor or even multiple sensors that do not employ sensory fusion. For example, in cases where the thermal identification of a rear-facing child seat is mixed with the seat upholstery and does not provide a movement signal, the distance measurement may be able to detect that something is in the seat with adequate reliability. However, in cases where passengers hold objects or are much larger than normal, an ultrasonic sensor will provide ambiguous distance measurements with a "view" similar to an RFCS. By the fusion method of this invention, the combination characteristics extracted from the inclined IR detectors and directed to "see" in different fields and from an ultrasound sensor can ensure the appropriate identification and output of an appropriate decision signal. In accordance with the present invention, the condition measurements are taken continuously and compared with the previous conditions to provide a current status profile. At least initially, the updates are compared with the initial conditions obtained in the start-up of the vehicle, and later, the comparison is with the conditions of the previous state. If the initial conditions indicate a recognized (or "valid") occupant classification, this condition will tend to prevail throughout the operation of the vehicle with the sensory algorithm always erring on the safety side. If the initial conditions indicate an empty seat, a "wakeful" mode ensures that passenger seat changes are detected during vehicle operation. An expectation mode can be provided while the ignition is turned off in order to drag less power and carry out only the periodic checks and the minimum maintenance functions required. The individual sensors will make the decisions themselves wrong under certain conditions but in unrelated, non-overlapping ways. The fused sensor approach of the present invention covers these failure modes to ensure reliable performance by requiring the analysis of many different signal characteristics before making a recognition decision. Ordinarily, to compensate for its own area of marginal performance, an individual sensor must become more and more sophisticated, raising costs. In contrast, the system of the present invention employs fused data from two or more inexpensive sensors, preferably three ultrasound (US) and two infrared (IR) sensors, thus achieving the required level of sophistication, even at a significantly reduced cost. In addition, in the dual sensory type operation, the self-diagnosis is improved by cross-correlation data from one sensor with data from the other. The correlation / cross correlation involves the comparison of time of occurrence, location, direction of movement, magnitude of the detected case, rate of change and the like, and includes the correlation of equal characteristics detected by the same or different sensors or different types of sensors, and the correlation of different types of characteristics from different sensors or sensor types. Although the preferred embodiment of the present invention utilizes active acoustic and passive thermal sensing for its inherent design, simplicity, and safety features, it will be understood that the present invention is not necessarily limited to the use of multiple sensors of the particular type disclosed. Although the selected sensors are not radioactive and do not exhibit electromagnetic, electro-optical or other harmful exposure to the occupants, it will be understood that other combinations of two or more sensors of different types for occupancy detection can be easily used to achieve simplicity and still high reliability, of the present invention by the sensor fusion method of the present invention. In any case, the sensors exposed here present any harmful exposure to the occupants; for example, the ultrasonic unit operates at a frequency well above the hearing range of humans and dogs. It should be understood that the present invention is not necessarily limited to being used in conjunction with an air bag system. It can also be used for security and protection purposes due to the combination of two different sensory characteristics such as the combination of thermal contrast and movement with acoustic distance and movement as shown herein, prove to be highly advantageous for its reliability and simplicity in a number of applications outside of a vehicle as well as other applications with a vehicle. It can be used as a security system for the property, both inside and outside a building. In the preferred embodiment of the invention, two inputs of infrared sensory units and three of ultrasonic sensory units are combined in a microprocessor circuit by means of a sensor fusion algorithm to produce an output occupancy status signal to the bag controller. air. The signal results from the preselected safety weight for the various parameters extracted from the two sensors (called characteristics), and after a fusion process which finally makes a decision that is extremely reliable. An empirical profile is provided, in the form of a look-up table, a value matrix, empirical relation (s), or an algorithm for a plurality of known objects (e.g., human occupant, empty seat, seat for child that looks backwards or towards the front, animal, packages, etc.) either as a generic interior profile or as a developed profile (empirically determined) for a particular interior. During the operation, the fusion processing compares the signals with an array of security values of known condition to produce a set of security-weighted values. By way of example, some selected IR features 14 and selected ultrasound features 13 are compared either directly or after fusion to arrive at a total safety level signal that results in the air bag deployment control that triggers The airbag will enable / disable the signal (or absence of signal). The signals emitted are compatible with AECM interfaces. The IR sensing unit advantageously includes dual sensing elements (typically each with six active segments) that look at different areas of the seat (e.g., the back of the seat and the seat itself.) Furthermore, the "view" of these elements sensor is directed towards parallel zones vertically oriented by means of one or more Fresnel-type lenses so that the "thermal movement" characteristics can be extracted from the change in the thermal identifications from zone to zone. The occupant performs the matrix fusion processing of sensors and the decision-making operation on the selected sensory outputs.The fusion matrix has weighted inputs to ensure reliability in the process of making a decision. with the "known" empirical condition and / or configuration data, calibration data, initial conditions and updated historical reference data are considered in the process to make a decision (issue an occupancy status signal) to the air bag controller either to suppress or not (enable or disable) the deployment of the air bag on the air side passenger in a collision. When merging features and feature vectors to make the decision, each individual parameter has only a partial effect, or "vote", on the last merger decision. The final decision is based on several conditions or states that reinforce that decision by requiring several phenomena or independent aspects of it to occur simultaneously. The fusion process of the invention produces the decision with greater reliability to a single sensor of phenomena or multiple non-fused sensors. In addition to carrying out the fusion decision making by multiple sensors, the process requires the periodic analysis of the sensory outputs to make sure that all the sensors work properly. In addition to normal electrical condition checks, the conditions of each sensor output are compared to the output of the other sensor to ensure that all sensors confirm their proper operation. In the different scenario where the sensory system completely fails due to the power failure, the failure of a component, or otherwise, the omissions of the airbag deployment system controller for the deployment condition in order to ensure the safety of the passenger. A diagnostic warning indicator of a fault condition can be provided on the vehicle's indicator board. All sensors of the present invention are co-located and preferably provided in a single unit at the junction of the upper part of the windshield with the front end of the performer to maintain the low cost of manufacturing and simplify the work of assembling the sensory system in a new vehicle or modernization in a previously assembled vehicle. In addition, the aesthetics of the vehicle is maintained by keeping the unit of the sensory system at a minimum size. Furthermore, due to the cross-correlation of the sensor inputs and sensor fusion, it is not necessary to separate the US from the IR, for example, it is not necessary to put the US on the instrument panel and the IR backwards, above the passenger. Having two or more sensors in the fusion mode improves the autodiagnostic correlation between the two, but if there is a failure of one but not the other, even in scenarios where a minimum or no signal of the failed sensor is expected, they will still be omitted some of the expected characteristics and the analysis and fusion will identify the faulty sensor. For example, if the US indicates an occupant, the IR can be received. If this does not indicate any occupant, then a potential sensor malfunction is indicated. If there are some characteristics of the IR, that is to say IR of weak signal, then the IR may be working but it is not clear what is in the seat until other characteristics received are analyzed by means of the algorithm of the fusion process of the invention. Although a fixed sensory system with tilted lenses is shown (for the IR), a mechanical path scanner can be used when mounting one or more sensors on a moving element. Similarly, although a fixed US transducer and receiver that uses the pulse to oscillate or receive the sensor is shown, a separate transducer and receiver may be employed. The acoustic signal profile can be configured inside for its maximum or narrowest coverage focused on a specific area. The IR sensor can be an uncooled electrical device that responds to IR radiation from the near-far IR (2-12 micrometers wavelength), and the US can be an electrostatic type sensor with a typical frequency range of 40 KHz to 150 KHz. The typical field of vision will be approximately 30 ° x 34 ° for the IR and 20 ° to 30 ° (direct or decentered conic) for the US. The US is highly immune to interference because the impulse echo must be received within a pre-selected time window as valid. The US beam can be asymmetric for better coverage. A separate IR sensor can be added to the search-oriented unit at the central (middle) location of the passenger. Without additional hardware, the system of the invention can automatically oscillate to "on" to measure the interior temperature of the vehicle in which it is installed and sends a signal to automatically adjust or cause the cooling fan to operate as long as the indoor temperature exceeds a preselected maximum value (selected design). Additionally, the system can automatically measure, in "power rise", the characteristic "identification" inside the particular vehicle in which it is installed, and when comparing these values with predefined reference tables embedded in the ASIC, it determines which type of platform is installed in, for example, a car or truck. It can then transmit the type of vehicle identification to the body controller, thus automatically verifying proper and proper functioning at the final installation / assembly point.

The ASIC of this invention allows several additional features to be optionally incorporated into the sensory system of this invention as desired. This includes: 1) Central Passenger Occupancy Detection (CPOD) that uses an additional IR sensor and lenses to detect occupancy of the center seat; 2) Four Quarter Temperature Control (FQTC). This system replaces the solar sensory and environmental control unit currently used. It not only controls the vehicle interior temperature, but also allows the automatic selection and control of one to four quadrants of directed HVAC (allowing up to four individual interior temperature settings): 3) Passive Thief Dispensing (PTD). The automatic temperature control sensors can be used to detect the presence of a person and the vehicle and through communication with the body controller can decide whether the entry was appropriate or not, ie a (appropriate) gain entry or no (inappropriate entry); 4) Near Object Detection Sensors (NODS). This system uses a microwave radar of extremely low power which can be mounted behind a plastic cover (flashlight or back), and adapted to detect objects within a preselected field of vision.

The FQTC is similar to the occupant sensor, and uses a "multi-aperture" lens to facilitate motion detection. In addition, the sensors are effectively "multiplexed" in the central network processor where the synchronization of samples, the duty cycle, and the sensory selection sequence are all programmable. The PTD uses thermistor bolometer (TB) detectors, instead of pyroelectrics, and is thus able to detect both the movement of a hot object and is able to determine its approximate temperature. This PTD implementation is configured electronically to provide continuous or intermittent selected vehicle monitoring. The electronics (Signal Conditioner, Power Regulator, Motion Detection Logic, etc.) are set to an extremely low current drain (less than 100 microamps) on the vehicle's battery during the "system power-on" state. safety, such as when the vehicle is not serviced with the ignition off. This configuration allows the active temperature monitoring of each zone while the car is in use. In addition, when the vehicle is left unattended, the sensor series is able to detect and report unwanted intrusion associated with vehicle theft or possibly a person hiding in the rear area of the seat. The NODS uses microwave radar (pulse) instead of the classic IR and Acoustic detection, but employs sensor fusion as discussed here. Microwave radar is used due to its ability to operate (invisibly) but protected from an external hostile environment when mounted in a rear fender or assembly location flashlight. This system has a reliable range detection of the order of 15 + feet. The hardware concept incorporates voltage protection, interface bus J1850 and one or more ASIC (s) for the control and implementation of the algorithm according to the principles of the invention. The specific frequency employed is in the range of from about 1.7 to 94 GHz. The sensory system and the methods of the invention coined the following properties: thermal identifications or contrasts coupled with movement to establish the presence of a hot object; and Acoustic identifications through wave propagation coupled with movement to establish the state of the object, that is, the distance from the location of the instrument panel or performer of the occupants, objects, empty seat, etc. and if they are animated or static.

Both sensory properties are required to meet the reliability requirements due to: 1) the need to inhibit the airbag when a rear-facing child seat is carried out more reliably through dimensional measurements; these are derived more reliably from the acoustic sensor. 2) the thermal conditions inside a vehicle change dramatically with the seasons, the weather, the interior of the vehicle, the clothes of the passenger, and the use of the driver. The use of an IR sensor can only lead to a higher proportion of falsely declared seat condition condition and, more importantly, failure to detect a present occupant. 3) the self-diagnostic capacity of the system requires sensory interaction / confirmation to improve its reliability. The signal processing employed in the fusion of multiple sensors of this invention is preferably implemented in one or more Application Specific Integrated Circuits (ASICs). In addition to the ASIC signal processor, a microcontroller provides the additional decision to develop the power and system control functions. The ASIC is an analogous device (A) and digital (D) of mixed signals, and can be a single combined A & D, or functions A and D in separate microcircuits. Signal conditioning, sensory signal detection, non-volatile storage, bus interface, status signal interface, and clock generation functions are carried out. The processing of parameters of the fusion matrix and the security weighting is conveniently carried out in the execution of the software on the microcontroller or can be implemented using logical cable connection circuitry. The software can be implemented by an expert in the art following the figures as described in detail herein. BEST MODE FOR CARRYING OUT THE INVENTION The following detailed description illustrates the invention by way of example, not by way of limitation of the principles of the invention. This description will clearly enable a person skilled in the art to develop and use the invention, and describes various embodiments, adaptations, variations, alternatives and uses of the invention, including what is currently believed to be the best mode for carrying out the invention. Referring now to the accompanying drawings, Figures 1 to 7 illustrate a variety of occupancy scenarios to which the present invention is generally directed in its preferred automotive occupation detection mode. As shown in Figure 1, this embodiment of the invention comprises a series of sensors 1 mounted in the area of the upper part, above and slightly towards the center of the passenger seat 12 of the vehicle 14. As described in more detail below, the microprocessor controller, including an ASIC having a microprogram described herein, is conveniently located in the sensory unit assembly 1 mounted on the performer 16 or dashboard 28. The sensing unit 1 is connected to a control unit. conventional air bag 2, which in turn activates an air bag 4 in an appropriate shock situation detected. The system is conveniently energized by the auto battery 6, or alternatively by the alternator or a separate slow charge gel cell (not shown). The various possible scenarios are represented by way of example in the following figures. Figure 1 illustrates the passenger seat 12 occupied by an average adult person 8, while figure 2 illustrates an empty seat. Figure 3 illustrates the presence of a child 10 in a rear-facing child seat (RFCS) 11 mounted on the passenger seat 12. The RFCS will have an unusual thermal pattern as well as distance and vibration identifications due to the possibility of that the child may be partly obscured by the seat, thus masking the natural thermal radiation. Figure 4 illustrates an adult person holding a grocery bag 18, which will also have unusual sensory readings. Figure 5 shows the presence of a child 10 in a front-facing child seat (FFCS) 20. Unlike the RFCS, this FFCS scenario will have a thermal identification closer to normal for a young child as well as readings of normal distance and movement. Figure 6 shows the presence of a pet such as a dog 13. Depending on the size and activity of the pet, there will be variation in the thermal, movement, and distance readings and the rate of change thereof. Figure 7 represents an illustrative scenario of a passenger out of position (OOP), where a child 10 stands on the passenger seat and is held or supported against the dashboard. It could also be a passenger adjusting the radio, or looking outside the front windshield or with their legs or feet on the dashboard. In this scenario, the sensory system needs to determine the feasibility of deploying the airbag, which depends on the distance of the passenger to the location of the airbag. If the passenger is too close to the location of the airbag, the deployment of the airbag may not serve any useful purpose, and instead could damage the occupant in the process.

The sensor unit is advantageously located on the front of the performer. Comparing for example, Figures 3 and 7, if the sensors are located at the X and / or Y position compared to the most universal, the intersecting position of the width-inclined performer / windshield 16 of this invention, RFCS 11 or the OOP occupant 15 may obscure or overload one or more of the sensors upon contact with the surface of the sensory unit or being too close to it. This is also an argument for the co-location of the sensors. The seat can also be occupied by passengers of different sizes, such as a small child or a larger person. An occupant can recline in the passenger seat or sleep in the passenger seat without much movement and both cases will have unusual movement, distance and thermal identifications. Referring to Figure 8, there may be inanimate objects 17 of various sizes on the seat, which may or may not give thermal and / or movement identifications. An example of scenarios that highlight strange or false signals include: a hot pizza box, a warm bottle, or a can of cold drink or frozen food. In addition to these scenarios, weather and shade conditions can affect the interior environment of the vehicle, especially the interior temperature of the vehicle. On a hot summer day, the passenger seat will be extremely hot after the vehicle has been placed close to the sun, and this condition can affect the sensory readings. In addition, driving along a branched highway can lead to thermal oscillation, which could simulate motion identification, due to intermittent overshadowing and seat exposure. The present invention is not limited to the detection of the scenarios described above, since others can also be detected. Given this wide variety of external and internal conditions of an occupant, the present invention must be capable of detecting, discriminating (categorizing) and making a decision to allow, directly or indirectly, through a signal of occupancy status, the transmission of a airbag enable signal, or generate a disable signal to the airbag controller to maximize passenger safety in the event of a collision. In the preferred embodiment of the present invention, these scenarios are categorized into one of the following five Occupation States: Empty state (negative or "E" state), Occupied state ("O"), Inanimate Object state ("10"), status of Child Seat Looking Back (" RFCS "), and status of Occupant Out of Position (" OOP "). For the Vacuum state, 10 state, RFCS status, and OOP status detected, an air bag disable status signal will be sent or supplied to the air bag controller. For the Occupant status, an enabled airbag status signal will be supplied to the airbag controller or, in the case that the default condition of the airbag controller is to signal its deployment to the airbag, no interrupted status signal will be sent from the sensory unit to the air bag controller (or air bag). Other modalities may include more or less states with variation in the scenarios. The Occupant status is the state where the deployment of the airbag will improve the safety of a passenger in the event of an accident. The Occupy status includes the scenarios of an average adult person, a small child, a child in a child seat facing forward, a passenger holding a grocery bag, a child standing in some positions, and the like . Note that in the standing child's scenario, the airbag will unfold if the child is far enough away from the deployment location of the airbag to allow an effective and non-damaging deployment of the airbag. The airbag will not deploy if the child is too close to the deployment location of the airbag, since the deployment of the airbag could harm the child by hitting it back towards the seat. The same consideration applies to OOP status, a passenger placed too close to an airbag can be detected to prevent a harmful deployment. Alternatively, a status signal allocating "slowly unfolding" or "partially inflated" information may be sent to the air bag controller, which may selectively deploy one or more bags with less than full charge, or characteristics of fast or slow inflation, for example, "hard" inflation vs. "Soft inflation" from single or multiple airbags. Typically, it is desirable to disable the airbag in the RFCS state, the Empty state, the OOP state, and the status of 10, for example, by sending an interrupted state signal or interrupting the deployment signal from the controller. airbag. It is especially important in the RFCS scenario that an airbag does not deploy in the event of an accident. The deployment of an airbag that hits the back of a child's seat facing back could throw the child and the seat backward, possibly harming the child in the process. In the case of a Vacuum or 10-state, the deployment of an airbag in the event of an accident ordinarily serves no useful purpose, and only adds to the repair cost the re-installation of a new airbag. in the vehicle. However, the system of the invention is diverted towards deployment to ensure the broadest level of security and reliability. In the preferred embodiment of the invention, the airbag controller is designed to miss the airbag deployment condition. For the appropriate states, such as the Empty state, the status of 10, the RFCS status, and the OOP status, the sensing system sends a disabled or interrupted status signal to the air bag controller. The present invention can also be adapted for use with a multiple-receptacle controlled air bag deployment system where the air bag is inflated by a number of receptacles up to the desired pressure. With this system, instead of sending a type of "on" or "off" signal, a quantitative series, or multiple parallel type of signal, can be transmitted to the air bag controller to indicate the desired pressure, or the number of receptacles to be released depending on the detected state. In order to recognize the different scenarios and conditions, this modality uses two types of sensors, an infrared sensor ("IR") and an ultrasound sensor ("US"). The infrared sensing unit used in this embodiment is a type of commercially available thermistor of infrared sensor unit, and there are preferably two or more detector elements, each having up to six segments, contained within the infrared sensor unit to enable sensory detection in or from two different regions. Although the pyroelectric and photovoltaic types of infrared sensors can also be used, the sensor thermistor type currently provides the best cost / performance ratio. In the presently preferred embodiment, the infrared detectors detect the objective areas continuously with an interrogation period of between about 2 Hz and 10z. The ultrasound sensor used in this mode is a commercially available ultrasound sensory circuit package where the frequency and the ultrasound pulse can be controlled externally. The sensor operates in the ultrasonic range above the hearing range of humans and animals such as dogs, and the typical frequency ranges are from 40 KHZ to 150 KH: frequency selection is determined by requirements such as losses acoustics, range, power, cost and size of the transducer. For example, the attenuation and absorption of air by the seats and clothes increases with frequency; however, the detection range required here is short, and as a result, the highest end of the frequency range can be selected. The higher frequency also provides the advantage that a small transducer source (detection element) can be used. In the currently preferred mode, the interrogation period varies between 2 Hz and 20 Hz during the actual operation depending on the quantity or quality of information needed. A series of three US sensors, colocalized in a single unit in the position of the performer of FIG. 1, is preferred. FIGS. 9A-9C are enlarged views of the sensory unit of the present invention shown in its place in the performer 16 of FIG. Figure 1. The sensors can be placed separately at different places, but in the preferred embodiment, as shown in Figure 9A, the series of infrared sensors 24 and the series of ultrasound sensors 26 are immediately placed one on the other in a single unit 22. Preferably, the infrared sensor has two or more detectors 21a, 21b separated by a vertical deflector 19 and covered by Fresnel lenses of multiple elements 23. Each detector 21a (Dl) and 21b (D-2) observes different portions of the seat, 21a observes the rear area of the seat 12b, and 21b observes the area of the seat 12a (see figure 10), through, in this example, two rows of Fresnel lens elements, FLa and FLb, which form a c set of lenses LS-1 and LS-2, respectively. Each row in this example has six individual lens elements 50a, 50b ... 50n, which observe the corresponding zones 50a, 50b ... 50n on the seat as seen in Figure 10. The fields of view of the row of lenses FLa are superimposed on row FLb, but individual zones 50a on 50n do not overlap. The deflector 19 is generally directed in the seat belt when properly used by the passenger, as shown by the arrow Q in Figure 9B. Figure 9B is a longitudinal schematic cross section of the IR sensor 24 along the line 9B-9B in which respect the horizontal angle? it can be 0 °, preferably varying from about 5-45 °, with 10-30 ° being preferred. Figure 9C is 24 taken along line 9C-9C of Figure 9A. It shows the orientation generally faceted 50a ... 50b. In the alternative, the elements can be staggered with respect to each other. Fresnel lenses allow the signal strength of a signal source from half of the zones to pass through them completely. However, as the signal source moves towards the edges of the zones, Fresnel lenses proportionally reduce the resistance of the signal passing through them. Although these sensory units can be placed in a number of places on the vehicle, it is preferred to place them on the performer 16 above the passenger seat at the junction of the upper part of the windshield with the front end of the performer as seen in the figure 1. The sensory unit can also be placed on the dashboard directly in front of the passenger seat or on pillar A on the passenger side. It is anticipated that in the future, rear passenger seats can be equipped with airbag protection as well. In this case, a sensory unit placed in the front and above the passenger seat in the performer or B-pillar can be used to detect the occupation of the rear seat. Figure 10 is a top view of the passenger seat 12 and the sensory unit 1. The passenger seat has a rear area 12b and a seating area 12a. Each area (rear and seat) is detected in multiple zones 50a, 50b ... 50n created by the Fresnel lens elements of the infrared sensor as shown in Figures 9A-9C. Note that the infrared sensor uses Fresnel lenses of the type in which each of the infrared detector's fields of vision is divided into, for example, five to eight zones. The infrared detector converts the photons (heat) to a change in the conductance of the detector, which results in a sinusoidal wave voltage when an object crosses each zone laterally. FIGURE IA illustrates a side view of the orientation of the two detectors 21a, 21b (FIG. 9) of the infrared sensor, observing the passenger seat 12. A detector 21b is oriented so as to observe the seating area 12a while the other detector 21a observes the rear part 12b of the seat. In addition to receiving the thermal identification data by zone, each infrared detector detects the lateral movement of the occupant or object crossing the zones 50a ... 50n is this designated area 12a or 12b of the passenger seat. By combining the data from the two infrared detectors, the "longitudinal" movement of the passenger can also be determined. By "longitudinal" movement is meant the movement of a passenger (for example, the hand of a passenger) that crosses from the area detected by a detector to the area detected by the other detector, and includes both anterior / posterior or frontal / backward movement (with respect to the vehicle) such as vertical or ascending / descending movement, or composite movement that has both anterior / posterior and vertical components. Figure 11b illustrates the area scanned by the ultrasound sensor 25 when it is directed to the seat, and the portions of the floor and dashboard 28. Referring now to aspects of the hardware, Figure 12 illustrates a schematic circuit for the preferred embodiment of the present invention. A specific application integrated circuit ("ASIC"), 30, is designated to receive data from the infrared detector series assembly 24 (SI) and the ultrasound detector series 26 (S2). The ASIC processes the data by controlling a commercially available microprocessor 32, and produces outputs to the Inhibit line on connector 28, the data bus of the computerized system on board the vehicle, J1850, on connector 27, and the line Diagnostics in the connector 26. The ASIC controls the transmission of ultrasound by modulating an "on" or "off" voltage through the connector 20 of the ASIC to the transistor, 34. The transistor is turned on in turn during a A short period of time to allow current to flow through the primary winding of transformer TI, which creates a current flow through the secondary winding of the transformer. The current flows to the transducer 27, which in turn transmits an ultrasonic pulse. The ultrasonic return signals are received by the transducer 27 and returned to the connector 19 of the ASIC. The infrared signals from the two IR detectors 21a, 21b (FIG. 9) of the unit 24 (SI) are received through connectors 22 and 21 of the ASIC. The incoming signals are amplified and filtered through capacitors, C5 and C6. The ASIC incorporates an algorithm in its hardware and software in the memory to process the signals and uses a commercially available microcontroller, 32, to do the calculations. The output signal of the resulting occupancy state is transmitted through the Inhibit line to the air bag controller. The ASIC also provides a diagnostic signal with respect to the integrity of the sensory system through the connector 26 of the ASIC for the air bag controller (ABC 2 figure 1) and the vehicle indicator board 28 (figure 1). In the event of a system failure, the air bag controller is missing the airbag enable state. The ASIC can receive inputs from the computerized system on board vehicle 3 (figure 1) through the J1850 data bus, the J1850, with respect to the various system conditions and environmental conditions that may allow the sensory system to consider certain environmental factors and vehicular conditions in their total calculations. The ASIC can also transmit its status or output to the standard computer on board the vehicle. The ASIC provides an oscillating clock signal to the rest of the instrument panel through connector 16. Figure 13 illustrates the functional description of the ASIC. Although the preferred modality is to have an ASIC microcircuit, the functions described can be contained in two or more ASIC microcircuits, such as analog in one and digital in another. The ASIC contains a Bus Interface J1850 40, Analog Outputs 42, a Non-Volatile RAM 44, a Digital I / O RAM 46, a Clock Generator &; Precision Oscillator 48, and a Synchronization & Control 49. The Digital I / O RAM 46 provides an AGC (automatic gain control) 51 and DIFFUSES AC Gain 53a, 53b and DC Gain 54a, 54b in infrared signal processing, and Ultrasound Control to an Ultrasound Transmission Control 56 in the ultrasound control through the connector 20. The Synchronization & Control 49 harmonizes data processing between an IR Feature Processor & FIFO 57, a US & Feature Processor FIFO 58, a US 59 Detection, a US 56 Xmit Control, and 1/0 Digital RAM 46. There are two infrared inputs and they are processed in the same way. The DC Gain 54a, 54b detects and accumulates the infrared signals to allow detection of the level by the Level Detector 60a, 60b. The fluctuation portion of the infrared signal is sent to the AC Gain 53a, 53b for motion detection and its sending to the Motion Detector 61a, 61b. The Level 60 Detector determines the amplitude and sends the information to the IR Feature Processor & FIFO 57. The AC Gain block 53 filters the jitter signal with the aid of a capacitor (C5 or C6) and sends the data to a Motion Detector 61, which sends the processed data to the IR Feature Processor & FIFO 57. The IR Feature Processor & FIFO produces IR Characteristics 62. The ultrasound signal is received through connector 19, amplified and filtered by a Gain & Filter 63, and sent to Detector US 59. Magnitude 64 and range 65 are extracted from the ultrasound data and sent to the US & Characteristic Processor. FIFO 58, which produces US 67 Characteristics. Both IR 62 Characteristics and US 67 Characteristics are sent to the Feature Combination Processor 66 to produce Fused Characteristics 68. For the case of multiple US sensors, each is directed slightly differently for give a wider or total coverage in the passenger compartment and signals of inner profile of somewhat different characteristic (see figures 24A, 25A, 27A and 27B) which can be compared, preferably the signals are multiplexed asynchronously, although with redundant circuits or duplicates could be processed synchronously. IR Characteristics 62, US 67 Characteristics, and Fused Characteristics 68 are sent to the Digital I / O Ram Block 46 for processing. The Digital I / O Ram 46 accesses a microcontroller through connectors 2 through 14 of the ASIC (Figure 12) to make the necessary calculations in order to process the data, and access the Non-Volatile Ram 44 for information . The results are sent out through the Bus Interface 40 and the Analog Outputs 42. In operation, the detection process is generally as follows: the incoming IR and US signals from the multiple sensors in an interrogation time period given are analyzed in terms of configurations (or characteristics) such as movement, frequency of movement, level of movement, temperature level, distance of objects, trends of increase or decrease, and so on. There is a set of characteristics for infrared signals and a set of characteristics for ultrasound signals. They combine ("merge") certain characteristics of each set to produce a third set of merged features. Each of the three sets or vectors is compared to a predetermined matrix of security levels and empirical relationships in order to determine a barely detected characteristic state. A characteristic state is one of the five possible states described above and is the state determined by the sensory system for this interrogation period. The barely detected characteristic state is compared to the current state. The current state is one of the five states above treated, and it is what the sensory system indicates is the real (almost present) condition of the passenger seat. If the barely detected feature state and the current state are different, a set of criteria is used to determine if the characteristic state should be returned to the current state. The current state determines whether or not a disabling or interrupting status signal should be sent to the air bag controller. The security levels, or the security criteria matrix, are determined as follows: security levels are data obtained from analytical and empirical studies of possible, known, predetermined passenger seating scenarios. Each such scenario is represented in the passenger seat under a variety of conditions, and the characteristics are obtained and analyzed. Some of the features are merged to obtain merged features. Generally, a security level is assigned to each feature and state combination. For example, in the currently preferred mode, five levels of security are used for most features. Some of the characteristics are not good indicators of some of the states for certain scenarios, so these particular characteristics have reduced or no security levels for those states. In more detail, from each scenario, the infrared characteristics and characteristics of ultrasound (or appropriate readings from additional sensors, or other types of sensors, if used). These characteristics from each scenario are compared with characteristics of other scenarios. After examining all the scenarios and their characteristics, values are assigned to each characteristic for each state. These values are called security levels, and they are assigned according to the resistance of the characteristic in the indication of the particular state. For example, in the case of a thermal level characteristic (quantitative quantity) from the infrared sensor, at five security levels from 1 to 5, with 1 being low security and 5 being high security, they can be assigned conveniently these possible characteristic values. After examining the thermal level characteristics of all the scenarios, the following observations are made: A thermal level of 1 (low thermal level) is a strong indicator of both the state of 10 and the empty state; at the same time it is a mean indicator of both OOP status and RFCS status, and a weak indicator of Occupant status. A thermal level of 3 (average thermal level) would be perhaps a high indicator of the state of RFCS and the state of OOP, a medium indicator of the state of Occupant, and a weak indicator of the state of 10 and the state Empty; The thermal level of 5 (high thermal level) would be a high indicator of Occupant status, a medium indicator of OOP status and RFCS status, and a weak indicator of Vacuum status and status of 10. After examining this characteristic, safety levels they are assigned according to the resistance of the indicators for each of the states. Through this process, all the characteristics are assigned security levels. Note that some of the features can be combined ("merge") to provide additional information about the scenarios and the security levels are assigned to the merged features as well. Conceptually, these levels of security are placed in a two-dimensional matrix with rows and columns, the columns being the characteristics or characteristics merged and the rows being the states. This matrix is referred to as the security criteria matrix. By examining all the characteristics and scenarios, empirical relationships between the levels of security developed from the combinations of characteristic and state can be deduced, and sets of empirical formulas can be derived to convert safety levels into probability values for each one of the states. More specifically, in empirical studies all related characteristics are collected and analyzed for that state. The interrelation (s) of the security levels for the characteristics are analyzed to determine how they relate in order to produce a high probability value for a particular state. From this examination, the empirical formulas are determined for this state. Then, using this set of empirically derived formulas in real (real time) scenarios, a probability value for the state is obtained. A set of formulas is derived for each of the states. A matrix of safety criteria and sets of empirical formulas for each vehicle model is developed due to the variations in the interior area and the configuration of the passenger seat for each of the vehicles. In Figure 14, a signal processing functional block diagram is illustrated for the preferred embodiment of the present invention. The infrared unprocessed data from each of the detectors 21a, 21b (FIG. 9) of the Infrared Sensor 24 (Unprocessed Data IR 1 70 and Unprocessed Data IR 2 71) are processed through the Infrared Feature Processing 74, the which produces an Infrared Feature Vector (A ') 76. Similarly, the Unprocessed Ultrasound Data 75 from the Ultrasound Transducer series 26 is processed through the Ultrasound Feature Processing 77, which produces a Vector of Ultrasound Feature (B ') 88. The Ultrasound Transducer can also transmit an ultrasonic pulse through the Synchronization & Ultrasound Transmission Impulse Control 87. A subset of the Infrared Feature Vector (A ") 78 and a subset of the Ultrasound Feature Vector (B") 79 are processed through a Fused Feature Processing 80, which produces a Fused Feature Vector (C) 81. These three vectors, the Infrared Feature Vector, the Ultrasound Feature Vector, and the Fused Feature Vector are processed by the Detection Processing 82, which produces a Feature State ( D ') 83. The State of Feature is processed by Decision Processing 84 with entries F "from a Diagnostic Controller 86, and the State of Feature is evaluated to determine a Current Status (E') 85. Depending on the Current Status, a status signal can be sent to disable or otherwise control the airbag, as shown to the bag controller of air The Diagnostic Controller 86 also indicates through the F1 system the health of the sensory system for example its good or bad function, and in the latter case the airbag is enabled. The feature sets are extracted from the signals for the given interrogation period. In Figure 15a, the Infrared Feature Processor 74, the unprocessed infrared data is digitized by a Digitizer 100 with reference to a Gain Calibration Data 101 obtained at the vehicle start-up and stored in the Memory 102. The Data Gain Calibration is used to calibrate the sensory readings. From this digitized unprocessed data, the frequency of the lateral movement of the object or objects in the passenger seat is extracted and calculated by a Frequency Processor 104 to obtain a component of Lateral Motion Frequency IR 1 106. As of same digitized unprocessed data, the thermal level of the object in the passenger seat is converted to one of the predetermined levels by a Comparator 108 to obtain an infrared thermal level component 1 110. The predetermined levels are levels that group correspondingly the Analog signal values in a set of discrete n-equal levels. This component is compared against the previously obtained thermal levels stored in the Memory 112 by a Temporary Processor 114 to determine the trend of the thermal level (increasing or decreasing thermal level), and produces an Infrared Thermal Temporal component 1 116. The unprocessed data digitized they are also filtered by a Pre-Filter 118 to improve the property of movement of the data, and the data is compared to predetermined levels of movement by the use of a Comparator 120 and an Infrared Lateral Motion Level component 122 is determined. The component is compared by a Temporal Processor 126 against the previously obtained movement levels stored in the Memory 124 to determine the movement level trend, a Temporal Infrared Motion 128 component. The unprocessed data from the second detector is processed from the same way to get a Level component of Side Movement IR 2 130, a Temporal Component of Side Movement IR 2 132, a component of Thermal Level IR 2 134, a Temporal Thermal component IR 2 136, and a component of Side Movement Frequency IR 2 138. The levels of movement of the two infrared detectors are correlated by a Motion Correlator 140 in order to determine a component of Longitudinal Movement Level 142, which shows any longitudinal movement of the occupant. The longitudinal information obtained from each detector is contrasted against each of the others to obtain an Infrared Differential Longitudinal Movement Level component 144, which is significant when there is movement by one detector but not by the other detector. This component is compared by a Temporal Processor 148 against the previously obtained components stored in the Memory 146 to determine the movement level tendency or a Temporal Infrared Differential Movement 150 component. The frequency of the longitudinal movement of the occupant is calculated by a Processor of Frequency 152 to obtain an Infrared Differential Movement Frequency component 154. The Infrared Feature Vector (A ') 76 is comprised of the infrared components described above, although only features 106, 110, 128, 154, 132, 134 are used and 138 to form the subset of IR Feature Vector A'2, 78. Referring now to FIG. 15b, which illustrates the Ultrasound Feature Processor 77, when an ultrasound pulse is transmitted to the objective area, the ultrasound transducer can receive several ultrasonic returns shortly after the impulse bounces in v several objects. These returns are digitized by means of a Digitizer 160 with reference to the Ultrasound Calibration Data 163 obtained at the start-up of the vehicle and stored in Memory 162. Each of these returns will have a point in time where the return starts first, called an edge, which is detected by an Edge Detector 164. And each of the returns will have a point in time where its amplitude is at the highest level (or peak level) and this point in time will be detected by a Peak Detector 166. The amplitude is compared to predetermined levels by a Comparator 168 to obtain return levels. From the edge level time and the peak of the returns, the Absolute Ranges 170 (or distances) of the objects of the sensory unit are determined. The first return of the transmitted impulse normally indicates the object of interest in the passenger's seating area and is the component of First Level of Return 176. The trend (increasing or decreasing) of the First Level Return component is the Proportion component of Change of the First Return Level 174, which is determined with reference to previous return levels stored in Memory 172. The Absolute-First Return 178 component is the absolute distance of the first object from the sensor. The movement rate of all the returns from an impulse is the Range Movement component? 80 found when using a Differentiator 102, and the movement rate of the Range Movement component is the Rate Movement Change Ratio component. 184 found when using a differentiator 186. The Range Movement shows the radial component of movement and vibration of an object. The trend of the Range Movement, faster or slower over time, is the Temporal Component of Range Movement 188 determined with reference to the previous range movement, values stored in Memory 190 and when using a Temporary Processor 192. The frequency of the Range Movement is the component of Range Frequency of Movement 194 determined by a Frequency Processor 196. The relative values between the returns are determined by a Rank Correlator 198 to find the components of Relative Range Values 200, the corresponding levels or the components of Relative Range Levels 202, and the trend of the Relative Range Levels or the Relative Range Level of Change component 204, which is determined by a Differentiator 206. Relative range level components tend to to indicate how many objects change in relation to each other and can indicate the movement of the object of i nterés. The range movement components indicate if there is a constant frequency of movement, which would tend to indicate an inanimate object, for example a vibration or trepidation, or if there are random movements, which would tend to indicate an occupant. The Multi-Path Triangulation component 208 is where the ultrasonic pulse bounces off several objects before being received by the transducer, and this value is compared by the Range Correlator 210 with the Range Calibration Data 162 obtained at the start-up. vehicle. This component is useful in determining if there is clarity in the scene being scanned. If the value of this component is low, it tends to indicate clarity in the scene and a corresponding high security in the scan. If the value of this component is high, it tends to indicate confusion in the scene and a corresponding low security in the scan. The Air Temperature 212 is obtained from the fact that the air is denser at a lower temperature than at a higher temperature, and there is a faster rate of return of the signal at a lower temperature because it is transmitted to a lower temperature. through denser air. The Ultrasound Feature Vector (B ') 88 is comprised of all of the ultrasound components described above, while the ultrasonic feature vector subset comprises only features 170, 178, 188, 194, 200 and 208. Now, referring to block C in FIG. 16, Fused Feature Processing 80, a subset of the Infrared Feature Vector (A ") 78 comprises the Differential Motion Frequency component 1.2 IR 144. , the IR 1 106 Side Movement Frequency component, the IR 2 138 Side Movement Frequency component, the IR 1 110 Thermal Level component, the IR 2 134 Thermal Level Component, the IR 1 12 Lateral Movement Component. , and the Temporal Side Motion Component IR 2 132. A subset of the Ultrasound Feature Vectors (B ") 79 for this embodiment comprises the components of Absolute Ranges 170, the Absolute-ler Range component. Return 178, the Trajectory Multiple Path Triangulation component 208, the Relative Range Value 200 components, the Temporal Motion Component of Range 128, and the Range Movement Frequency component 194. The two subsets are used to extract components from merged characteristics for the Fused Feature Vector (C) 81. The Infrared Spatial Frequency Components 300 are sets of distance, frequency, and levels of the objects calculated by the Spatial Correlation Processor 302, which determines the distance, the frequency of movement, and the size of the objects detected by the two sensors. The component of Absolute Surface Temperature IR 1 304, the component of Absolute Surface Temperature IR 2 306, and the component of Absolute Surface Temperature Absolute IR 308 and, respectively, the temperatures and the temperature distance found when using the Temperature Processor 310. Like the cross correlations, the Infrared / Ultrasound Motion Level Correlation component 312, the Infrared / Ultrasound Motion Level Temporal Correlation component 314, and the Infrared / Ultrasound Frequency Correlation component 316 are levels of motion, the movement trend (slower or faster), and the movement frequency as determined by the Correlation Processor block 318. Note, all components of the Fused Feature Vector (C) 81 are calculated by merging the features of both infrared and ultrasound sensors. Referring now to Figure 17, which illustrates the Detection Processor 82, each of the vectors is processed by its own respective feature security processor and security criteria matrix. Feature components are processed individually and some of the feature components are merged for processing. Referring first to the processing of the Infrared Feature Vector, the individual or fused components of the Infrared Feature Vector (A ') 76 are processed by an Infrared Feature Security Processor and Fusion Infrared Feature 400. The array was stored in the Memory 402, which is modified by the previously processed data stored in an Intermediate Memory of History 404. This process produces a Security Matrix and Detection of Infrared Feature 406, which is processed by a Fusion Detection Processor of ler. Infrared Level 408 to produce an Infrared Detection Decision Security Vector 410. The Infrared Detection Decision Vector / Ultrasound 412 and the Ultrasonic Detection Decision Security Vector 414 are produced in the same way with their blocks of processing, history buffers, and respective memories. The Detection Fusion Processor 416, with reference to the previously processed data stored in its History Interim Memory 418 and through the use of empirical formulas and relationships between the three detection decision security vectors (described above), produces a Feature State (D ') 83. A Feature State is one of the previously mentioned states: Occupied state, Empty state, RFCS state, OOP state, and state of 10. The three vectors, Vector of Infrared Feature (A ') 76, Ultrasound Feature Vector (B') 88, and Fused Feature Vector (C) 81, are used to produce a Feature State (D ') 83 as follows: using the Feature Vector Infrared as an example, leave the Infrared Feature Vector =. { IRF1, IRF2, IRF3, ..., IRF14} , where each of the IRF # represents a component, and where the Infrared Feature Vector has fourteen vector components (as shown in Figure 18). In the processing of the Infrared Feature Vector components, the security processor (e.g. Infrared Feature Security Processor and Fusion Infrared Feature 406) refers to a matrix of security criteria (e.g. Infrared Security 402), which is an empirically developed data through examination under various conditions and scenarios, as described above. The security criteria matrix contains security levels, which can and usually are modified by previously processed data. The security levels indicate the probability of the states for the given feature component values. For each relevant characteristic component or merged feature component, there is a set of security levels for each state. For example, referring now to Figure 18, for a particular Infrared Feature Vector Component ("IRFi") and states, an IRFi component value of 5 has an associated security level of 1.3 for the RFCS state, an security level of 1.3 for the OOP status, and a security level of 0 for other states. For an IRFi value of 9, it has a security level of 3.3 for the state of 10 and 0 for other states. Security levels can be modified by previously processed vectors stored in the History Intermediate Memory, and can be modified taking into account the environment and other changes. For example, if recent history shows that the interior of the vehicle has changing thermal characteristics, for example when starting the vehicle in cold weather with a warmer or more complete jet and then maintaining a consistent and warm temperature, the criteria matrix Security is adjusted to take this change into account. Since there is a higher total thermal level in the vehicle, a indicates the presence of occupants or their movements. In this way, over time, the level of security for each of the states may vary. Figure 19 shows a graph of the security level for a state of a particular vector component that changes over time. There are also security levels of merged features, where two or more vector components can indicate levels of security for the states. For example, with reference to Figure 20, an IRF5 value of 1.2 and an IRF1 value of 1.2 would result in a high safety value for the OOP status and 0 for the other states; an IRF5 value of 3 and an IRF1 value of 1 will have a security level of 0 for all states; and an IRF5 value of 2 and an IRF1 value of 3.3 will have a low security value for the RFCS state and 0 for other states. For each feature vector, there is a number of these possible fused vector components and their associated security levels. The output of the feature and merged feature processing block is a matrix, called the detection and security matrix (eg, Infrared Feature Detection and Security Matrix) shown graphically in Figure 21. Note that a merged vector can merge two or more characteristic vector components. The Security and Detection Matrix of Infrared Feature 406 (Figure 17) is input to the Fusion Detection Processor of ler. Infrared Level 408. In the previous stage, security level calculations provide each individual Infrared feature or merged features with its own detection "decision". These individual decisions are now Factorized together by the state in the relations and functional formulas derived empirically, as described above, that is: IR Security (RFCS) = Function of. { IRFl (RFCS), IRF2 (RFCS), ..., IRFn (RFCS), IRF3, 4, 5 (RFCS), IRF1,10,11 (RFCS), IRF8, 12 (RFCS), ...}; Security IR (Busy) = Function of. { IRF1 (Busy), IRF2 (Busy), ..., IRF8, 12 (Busy), ....}.; IR Security (OOP) = Function of. { IRFl (OOP), IRF2 (OOP), ..., IRF9, 11 (OOP), ....}.; IR Security (IA) = Function of. { IRFl (IA), IRF2 (IA), ..., IRF8.12 (IA), ...}; and Security IR (Empty) = Function of. { IRF1 (Empty), IRF2 (Empty), ..., IRF9, 11 (Empty), ....}. . Each of the above functional relationships will produce a value that indicates the level of security (or probability value) of the associated state. The output of this process is a vector, called the detection decision security vector (for example, the Infrared Detection Decision Security Vector 410 in Figure 17), where each state has an associated security value. An example of the Detection Decision Security Vector is: the Infrared Detection Decision Security Vector = OOP state: 0.02, Empty state: 090, RFCS state: 0.04, 10: 0.0 state, Occupied state: 0.20 \ . In the same way, the Vector of Ultrasonic Detection Decision Security 414 is produced from the Ultrasound Feature Vector 88, and the Infrared / Ultrasound Detected Decision Security Vector is produced from the Merged Feature Vector 81. Continuing in relation to Figure 17 , there are three independent detection decision security vectors, Infrared 410, Infrared / Ultrasound 412, and Ultrasound 414, are input to a Detection Fusion Processor 416, which produces a Feature State 83 (see also FIG. 14). The manner in which the decision of the Feature State is reached includes weighting functions associated with each security vector and the recent decision history weighting stored in an Interim Memory 418. For example, in the case of an RFCS, a Based on analytical and empirical studies, we found that the infrared characteristic is a "weak" indicator, the characteristic of ultrasound is a "strong" indicator, and the combined infrared / combined ultrasound characteristic is a "moderately strong" indicator. With these three characteristics, more weight will be applied to a declared RFCS state of ultrasound, less weight will be applied to the declared RFCS state of merged feature and even less weight to an infrared declared RFCS state. In this mode, all three detect decision vectors, the IR Detection Decision Security Vector, the US Detection Decision Security Vector, the IR / US Detection Decision Vector, are weighted and combined to produce a single vector with a corresponding security value for each of the states. The state with the highest security value is selected as the characteristic status. To summarize the processing of the Feature state, by using the feature vector and the matrix of security criteria set to time as input, the processor performs essentially a look-up table function for the security levels on each vector component. or fused vector component for each state. In this way, decision making is made independently at the infrared, ultrasound, and infrared / ultrasound characteristic level. Also, in this process, some features do not provide information about some of the states because these characteristics alone are not reliable to make correct decisions for those states. Although some features are not reliable to make correct decisions for some of the states, in combination, these characteristics are reliable to cover all states, and this is the power behind the use of fusion of multiple features from different sensors. Note that the preferred process described above involves first extracting features from unprocessed sensory data, production after merged features, associating levels of security with merged characteristics and characteristics in order to produce security levels for predefined states, and the determination of a characteristic state from the security levels of the states. This process uses merge at the characteristic level and at the detection level; It is not a simple error correction routine. Other methods of fusion can be employed within the principles of the present invention. An algorithm may also be used under certain circumstances to merge the raw sensory data before any characteristic is extracted. An algorithm can be used to extract the characteristics and produce a characteristic state from all the extracted characteristics. Similarly, an algorithm can be used to extract the characteristics of each sensor, produce a state for each sensor, and merge the states to produce a characteristic state. In other words, the data fusion can be done at the raw data level, feature level, decision level, or combination thereof, and any of the above algorithm or combination thereof can be used for the present invention.

The preferred embodiment uses a fusion combination at the feature level and at the detection level, and empirical comparison studies show that this preferred combination provides superior accuracy in detection and discrimination for a highly reliable decision. Referring now to Decision Processing 84 (E) in Figures 14 and 22, the Segu Processor L < Decision Aa 500 compares the State of Feature (D ') against a Current Status 502, the Status Change Criterion 503 stored in the 504 Memory, an Intermediate Memory of History 506, and an Intermediate Memory of the Health Status of the System 508 The Current State is the state condition as determined by the sensory system, ie, what the sensory system indicates is the state of the passenger seat, and the corresponding signal to maintain an enabling or disabling signal to the controller. airbag. If the currently detected Feature State is the same as the Current State, the Current State does not change and the History Intermediate Memory stores the Feature State. If the Feature State is different from the Current State, the Decision Security Processor determines whether the Feature State should become the Current State. For the Current State to become the State of Feature, it must satisfy the Status Change Criterion stored in the Memory, which is a set of predetermined criteria to ensure the highest level of security and reliability in the decision to enable or disable the deployment of the airbag. Generally, the set of predetermined criteria requires that more confirmations be made before switching from a deployment status to a non-deployment state, and fewer confirmations are made going from a non-deployment state to a deployment state. The Decision Security Processor also searches the history (from the start-up of the vehicle) of the Current States stored in the History Intermediate Memory and considers which Current State decisions have been made and how often the Current State has changed . The History Intermediate Memory is updated by the Decision Security Processor. In addition, a Diagnostic Controller 510 verifies the integrity of the sensory system and updates the Interim Memory of the System Health Status. The Diagnostic Driver provides a 512 System Health indicator to the air bag controller and to the vehicle indicator board. In case of system failure, the airbag controller is missing the airbag deployment condition, for example, by not sending an interrupt to the airbag controller. The Decision Security Processor verifies the Interim Memory of the Health Status of the System and the other conditions of the system to ensure that the sensory system is functioning properly. As an example of a state change decision process, if the Current State is the initial empty state with the corresponding signal to disable the airbag and the State of Feature is the Occupied state, the Decision Security Processor will verify the Interim Memory of the State of Health of the System to ensure the appropriate integrity of the system. It will also check the History Intermediate Memory to see how many of the previous consecutive periods have the Occupant status as State of Feature or how often the Current State has changed. The Decision Security Processor will change the Current State from the Empty state to the Occupant status if, for the last two periods, for example, the Feature State has been the Occupied state. On the other hand, if the Current State has been the Occupy state, it will take much more than two periods to change the Current State from the Occupied state to the Empty state. If the current state has changed quite a few times previously, it will be increasingly difficult to change the current state of occupant to empty state. This is because the preferred modality diverts decisions taking into account the change of state towards security. Figure 23 shows, in the case of detecting an occupant facing forward and allowing the airbag to unfold, while inhibiting deployment if an RFCS is detected, that the dual sensory system of the invention provides functional reliability very high Reliability, R, of .98 (98%) or greater is obtained using sensor fusion even when the probability of PD detection for Sensor 1 is as low as .3 and the probability of false detection, PFA, is so high as 10"4 (R of .27), the only Sensor 2 has a PD of .99 and PFA is of 10 ~ 6. The AOS of this invention can even recognize the vehicle in which it is found by measuring the relative position of the module and the interior attributes of the vehicle Figure 24a shows the actual measurements carried out by the AOS system described above in a Chrysler LH vehicle The range indications show the acoustic returns with real time reference from the test vehicle, the arrangement of which is shown in Figure 24b. Figure 25a shows the real measurements carried out on a pick-up truck Dodge 1989 layout shown in Figure 25b. Table 1 below shows the actual synchronization values and measured by the AOS system. These results show a signal margin of 1060 μs in the brand of IP measurement, 250 μs on the seat position mark and 543 μs at the floor mark. The total time difference is 1860 μ. With a time resolution better than 20 μs, the AOS has a large margin of signal processing when it identifies the difference between vehicles such as "" "" 15 a Chrysler LH and a RAM truck. The comparison of the indices of Figures 24a and 25a shows the unique identifications of the interior vehicle configurations by which the AOS of this invention can recognize the vehicle, and a normal state thereof. 20 TABLE 1 LH TRUCK We have measured several types of significant data to evaluate the potential performance of the OSA. These data show excellent signal and noise ratios (SNR) and a large margin of design performance of the sensor series. The signal-to-noise values and the resulting predicted performance are summarized in Figure 26. The Pd numbers in Figure 26 were calculated using the fused probability equation of 4 characteristics shown below. R?, 2,3,4 = R 1 + R 2 + R 3 + R 4 - R? (R 2 + R 3 + R 4) - R 2 (3 + R 4) - R 3 R 4 + R 1 (R 2 R 4 + R 3 R 4) + R 2 (R 3 R 4 + R 3) - R1R2R3R The individual probability entries to the equation were derived from the actual measurements and the worst case analysis. The examination conducted on typical IR detectors produced a SNR in the range of 12: 1 from a normal occupant in an 83 ° F vehicle. The ultrasound sensor produced a SNR of 16: 1 during the same type of test. By way of comparison, the return of the ultrasound sensor from a rear-facing child seat was measured with the RFCS both covered and uncovered with two wool blankets. The child seat was a Century mark and was placed in an Eagle Vision 1993. The uncovered child seat gave a SNR of 20: 1 while the seat covered under two blankets generated a SNR of 11: 1.

These signal indicia were shown in Figures 27a and 27b, respectively. These data indicate that the system of the invention can easily discriminate even between these two subtly different occupant states. The measurements reflected in Figures 22-27 were taken under static conditions in the laboratory. Assuming that under the worst case conditions, the signals would degrade by approximately a factor of 4, all SNR data are divided by 4. With only small gains in signal processing, the data was increased by a factor of 2. This small gain of signal processing does not include the use of any historical input or adaptive threshold in the detection process, which are standard techniques that can provide a substantially increased signal processing gain. Because this is a worst-case analysis, such adaptive and historical gains are not included. Using the performance numbers of the worst case adjusted system, the detection probabilities were calculated for each detection mode. The calculation assumptions used here were a simple envelope detection using fixed thresholds in a Gaussian noise distribution, as long as the AOS of the invention uses more sophisticated detection processes and has greater probabilities of individual detection to ensure adequate Pd under all condition. The probabilities of detection of individual sensor mode are shown in Figure 26, and were used to calculate the probability of fused detection shown in the right-hand column of Figure 26. For this analysis a lifetime of 15 million cycles is assumed. . The probability of false alarm for this analysis was set at one in a million cycles. The probability of false alarm will be reduced to an even smaller number when considering the historical and adaptive processing gains. Failure to include these gains shows the performance of the system in the worst case. Diagnostic reliability also benefits from the fusion of multiple sensors to almost the same extent as the detection benefits. As shown in figure 26When each sensory diagnostic probability is fused, the diagnostic probability of the resulting system increases. As was done for the detection analysis, the diagnostic probability numbers started as laboratory measurements that were adjusted downward for the worst case conditions, then adjusted for the gain of the worst case signal processing. These individual probabilities were taken from the noise iO - Gaussian and a false alarm ratio of one in 100 million cycles. For both PI (Instrument Panel or Dashboard) and top locations, operability was evaluated and examined. High occupant reliability and detection of the child seat facing backward from both the IP and top positions can be accomplished by the use of sensor fusion techniques of this invention. Both IR and ultrasonic sensory performances have been determined as independent locations, but as noted, co-localization of the frontal performer is preferred for the reduced cost and the incidence of sensory block noted above. Figure 28 is an isometric view of the AOS 1 of this invention as a single unit 22 in a suitably configured housing. Two IR sensors 24a and 24b co-located with three US 26a, 26b and 26c transducers are shown. Each one "looks" over its designated field through holes with appropriately sloped hole 600a-e as shown. DISCUSSION The sensory position in the upper part offers the system performance advantages over the mounting position in the instrument panel (IP). The position at the top is much stronger to block intentionally by the normal behavior of the occupant. In the top position, the relative geometry of the vehicle is measured much easier. This feature allows an AOS mounted on the top to measure the relative position of the IP, seat and floor, and determines the type of vehicle in which the AOS has been placed. With respect to this, it should be noted that it is not necessary to separate the US sensor (s) from the IR sensor (s), and it is preferred to co-locate them in a single unit of the upper part placed adjacently. to the junction of the windshield and the front end of the performer. The use of plural sensors, currently preferred is three US and two IR sensors (each with six segments of active element) and the fusion of sensors for the cross-sensory correlation of the characteristics that come from different phenomena allows the determination of the presence or absence of an object and categorization or classification as to the nature of the object, that is, if it is an occupant, and if so, discriminates between types, for example, RFCS vs.. FFCS, a human passenger, a box, an animal, etc. The automotive occupation system of the invention need not necessarily create per se a enable or disable signal. Instead, it generates a status signal as described above on which the microprocessor can act to enable, disable or otherwise control the relative display, deployment ratio and / or volume of one or more bag (s). ) of unfolded air (s). The sensor fusion techniques of the invention employ orthogonal features within the same sensor that do not correlate otherwise. The AOS system correlates the characteristics derived from different phenomenology through the different sensors; and shows the same characteristics of the uncorrelated sensors. The AOS of the invention searches in all the characteristics, using the linear transformation of some characteristics -30 processed in parallel before elaborating an occupation decision state including or not presence, location and categorization of the nature of the present object. It should be understood that various modifications may be made within the scope of this invention by one skilled in the art without departing from the spirit of the same. For example, memory and history buffers may be used to store the state decision for a predetermined period (for example zero to approximately 60 to 600 seconds, depending on the size of the memory supplied in the system, before a clash) in order to determine what the occupants did before or during the crash. Was a dog out of position, a passenger made unusual movements indicative of distractions or intrusions, etc.? This can be downloaded from moment to moment in a special memory in a "black box" of shock along with other vital operation data of the vehicle, fuel level, speed, acceleration / deceleration, change of direction, breakage, lights and / or windshield wipers lit, indoor climate and the like. The current state history (and the initial state decision) is saved in the RAM. When the airbag control module signals threaten, probably with a probable or possible shock (fired impact), these data are discharged into an EPROM in the AOS unit for post-shock extraction and analysis. The typical history of interest would be from the shock signal back in time from about 1/2 second to 30 seconds before the shock, and may extend through the event of the shock, as long as the sensors are preserved and functioning without harm These data can also provide the analysis of compartment damage during the crash. In addition, once the computerized system to 14 - board 3 (figure 1) produces a disable signal to the airbag, it can be at the same time of the signal an alert in the passenger compartment, for example, by means of a warning / reminder light on the dashboard or the activation of an electric bell, buzzer or a voice circuit with a warning message or warning that the airbag is deactivated. Accordingly, we desire that our invention be defined by the scope of the appended claims as broadly as permitted by the prior art and in view of the specification if necessary.

Claims (10)

1. NOVELTY OF THE INVENTION Having described the present invention, it is considered as a novelty and therefore the property described in the following claims is claimed as property. 1. A method for determining whether or not to deactivate a passenger passive restriction system of a vehicle as a function of a given current state value by comparing the measured signal characteristics with a predetermined set of safety values and empirical relationships, obtained by using various known occupancy scenarios and a set of status change criteria, comprising the steps of: (a) detecting occupancy characteristics of a particular passenger seat within the vehicle using a plurality of functionally associated sensors with said passenger seat and develop a set of corresponding electrical signals; (b) evaluating said electrical signals to determine a plurality of signal characteristics included in each of said signals; (c) combining certain of said signal characteristics to obtain a plurality of merged characteristics; (d) associating said signal characteristics and said merged features with the security values and the empirical relationships to determine a characteristic state value; (e) identifying the characteristic state value as the current state value in the state change criteria set that is met: and (f) generating a deactivated signal if said current state value is one of a predetermined subset of values for which said passive restriction system is deactivated. A method according to claim 1, characterized in that said predetermined set of state values includes values corresponding to an empty seat state, a occupied seat state, a rear-facing child seat state, a passenger-out state of position, and a state of inanimate object. A method according to claim 1, characterized in that said association step (d) comprises the sub-steps of: i) using predetermined security values in said signal characteristics and merged characteristics to produce (1) a security decision matrix of the safety matrix of the safety values for the signal characteristics of the signals of each sensor, and (2) a decision security matrix of the security values for said merged features; ii) use the empirical relationships to calculate a decision security vector corresponding to each of said decision security matrices; iii) weight each decision security vector in a predetermined manner to produce weighted vectors; iv) combining the weighted vectors to produce a resulting vector having state values from which the characteristic state value is selected; and v) status values corresponding to a backward facing child seat state, an empty seat state, an inanimate object state, and an out-of-position state. 4. A method according to claim 1, characterized in that said passive restriction system includes an air bag deployment system with a single receptacle. A method according to claim 1, characterized in that said passive restriction system includes an airbag deployment system with multiple receptacles capable of partially pressurizing an airbag to various degrees of pressure. A method according to claim 1, characterized in that said signal characteristics include indications of lateral movement and longitudinal movement as detected using multiple element Fresnel lenses to focus one detector on a rear part of the passenger seat and focus the other detector on a seat surface of the passenger seat. 7. An apparatus for determining whether or not to control a passive passenger restraint system of a vehicle as a function of a given current state value by comparing the measured signal characteristics with a predetermined set of safety values and empirical relationships, obtained by using various known occupancy scenarios and a set of status change criteria, comprising the steps of: (a) means for detecting occupancy characteristics of a passenger seat within the vehicle using a plurality of sensors functionally associated with said passenger seat and a set of corresponding electrical signals; (b) means for evaluating said electrical signals to determine a plurality of signal characteristics included in each of said signals; (c) means for combining certain of said signal characteristics to obtain a plurality of fused characteristics; (d) means for associating said signal characteristics and said merged features with the security values and the empirical relationships to determine a characteristic state value; (e) means for identifying the characteristic state value as the current state value if the state change criteria set is met: and (f) means for generating a deactivated signal if said current state value is one of a predetermined subset of values for which said passive restriction system is controlled, including the possible deactivation of said restriction system. An apparatus according to claim 7, characterized in that said plurality of sensors is selected from the group consisting of infrared sensors, ultrasonic sensors, weight sensors, microwave sensors, capacitance sensors, light sensors and laser sensors. An apparatus according to claim 7, characterized in that said plurality of sensors includes: i) a first infrared detector for generating a first unprocessed data signal; ii) a second infrared detector for generating a second raw data signal; and iii) at least one ultrasound detector for generating at least one third raw data signal; and wherein said means for the evaluation step (b) include: i) means for processing said signals, first and second, of raw data and to develop a first set of signals representing a first group of signal characteristics and defining an infrared characteristic vector signal; ii) means for processing said third (s) raw data signal (s) and for developing a second set of signals representing a second group of signal characteristics and defining at least one ultrasound characteristic vector signal; iii) means for selecting a subset of said first group of signals to develop a third group of signal characteristics that define a subset signal of infrared characteristic vector; and iv) means for selecting a subset of said second group of signal characteristics to develop a fourth group of signal characteristics defining at least one ultrasound characteristic vector subset signal; and wherein said means for the combining step (c) include: i) means for processing said infrared characteristic vector subset signal and said ultrasound characteristic vector sub-signal (s) to develop a signal of fused feature vector; and ii) means for correlating a first subset of said third group of signal characteristics with a first subset of said fourth group of signal characteristics and for developing a signal of infrared spatial frequency components; (iii) means for processing a second subset of said third group of signal characteristics with a second subset of said fourth group of signal characteristics and for developing a first absolute infrared surface temperature signal, a second absolute surface temperature signal infrared, and a differential infrared absolute surface temperature signal; (iv) means for processing a third subset of said third group of signal characteristics with a third subset of said fourth group of signal characteristics and for developing at least one infrared / ultrasound motion level correlation signal, a signal of temporal correlation of infrared / ultrasound movement level, and an infrared / ultrasound motion frequency correlation signal; and (v) where the signal of infrared spatial frequency components, the first absolute infrared surface temperature signal, the second infrared absolute surface temperature signal, the differential infrared absolute surface temperature signal, the signal (s) (en) infrared / ultrasound motion level correlation (s), infrared / ultrasound motion level correlation signal (s), and infrared motion frequency correlation signal (s) / ultrasound combine to form said fused feature vector signal. 10. An application-specific integrated circuit device for processing the input sensory signals received from the sensors adapted to detect the occupancy characteristics of a particular passenger seat within a vehicle, and to determine whether or not to deactivate a system Passive restriction of the passenger of a vehicle as a function of a current state value determined by comparing the measured signal characteristics with a predetermined set of safety values and empirical relationships obtained by using various known occupancy scenarios and a set of state change criteria, comprising in one or more microcircuits: (a) means for evaluating said input signals in order to determine a plurality of signal characteristics; (b) means for combining certain of said signal characteristics in order to obtain a plurality of merged features; (c) means for associating said signal characteristics and said merged features with the security values and empirical relationships to determine a characteristic state value; (d) means for identifying the characteristic state value as the current state value if the set of status change criteria is met; and (e) means for generating a deactivation signal if said current state value is one of a predetermined set of state values for which said passive restriction system is deactivated.