Real-time multi-sensor integrity monitoring through sensor fusion for navigation purposes
By using a single Kalman filter to evaluate sensor integrity through sensor fusion, the problems of sensor performance degradation and inaccuracy caused by spoofing are solved, resulting in a more reliable PNT solution.
Patent Information
- Application Number
- CN202380100305.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-17
- Publication Date
- 2026-02-24
Smart Images
Figure CN121569217A_ABST
Abstract
Description
Background Technology
[0001] In the field of navigation, sensor fusion can be performed using different types of sensors to provide a more reliable and accurate estimate of PNT (Position, Navigation, and Timing) solutions. Sensors commonly used to provide PNT information include GPS / GNSS receivers, IMUs (Inertial Measurement Units), SOP (Successive Opportunity Signal) receivers, precision clocks / oscillators, odometers, visual navigation systems, barometers, altimeters, etc. Sensor fusion algorithms typically rely on filtering, such as Kalman filtering or particle filtering, to combine noisy sensor measurements into an optimal "fused" solution. Filters create dynamic weights based on the input measurements and their associated uncertainties, but typically do not consider sensor integrity. Sensors impaired due to performance degradation, rejection, or spoofing can adversely affect the fusion solution and may lead to inaccurate results.
[0002] Therefore, there is a need for methods and systems that can provide accurate PNT solutions. Summary of the Invention
[0003] This invention provides a simplified overview of some concepts, which will be further described in the detailed embodiments below. This invention is not intended to identify key features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.
[0004] This document discloses a method and system for determining a localization, navigation, or timing (PNT) solution. In some embodiments, the system utilizes a single Kalman filter configured to compute multiple combinations of sensor information to simultaneously compute multiple PNT solutions. These multiple solutions include solutions from multiple independent sensors and different subsets of multiple sensors. The computed solutions are compared with a master fusion solution for integrity scoring evaluation. If it is determined that the deviation between the solution from an independent sensor or a subset of sensors and the master fusion solution exceeds a threshold, the solution from a subset of sensors without impaired sensors is selected as a new master solution to be reported by the system. Attached Figure Description
[0005] The above aspects and many incidental advantages of the invention will be more readily understood and fully appreciated from the following detailed description taken in conjunction with the accompanying drawings, wherein:
[0006] Figure 1 This is an exemplary system according to the present technology;
[0007] Figure 2This is another exemplary system based on the present technology;
[0008] Figure 3 This is an exemplary method for generating positioning, navigation, or timing solutions according to the present technology; and
[0009] Figure 4 This is a block diagram of an exemplary system according to the present technology. Detailed Implementation
[0010] Although illustrative embodiments have been illustrated and described, it should be understood that various changes may be made therein without departing from the spirit and scope of the invention.
[0011] In some embodiments, the system computes multiple PNT solutions without using multiple filters or algorithms to compete for computational resources. The linear algebra within the Kalman filter allows for the addition of additional data rows and columns using different combinations of sensor information. This slightly increases the complexity of the filter, but is still much less complex than using multiple independent filters. Different filters, such as particle filters, least-squares filters, Wiener filters, etc., can be used in different embodiments.
[0012] As described herein, a single sensor is defined as a single component that provides positioning, navigation, and / or timing information. Furthermore, sensors that improve measurement confidence or assist in direction finding, such as magnetic compasses providing heading guidance, barometric altimeters, or visual aids, can also be considered sensors. In some embodiments, each individual sensor is continuously monitored for a predetermined integrity metric (also known as severity impact) specific to that sensor. A single sensor capable of providing sufficient information to form a standalone PNT solution can also be considered a single sensor.
[0013] As disclosed herein, a subset of sensors (or a sub-combination of sensor performance) is defined as a grouping of components that provide positioning, navigation, and / or timing information. A subset of sensors is described by similar characteristics (visual, radio frequency (RF), physical, etc.), available data provided (positioning, navigation, and / or timing), interference opportunities (magnetic, optical interference, RF interference, multipath, none), and statistical performance analysis.
[0014] Individual sensors and subsets of the entire sensor suite offer different sensor performances. Each sensor performance is scored using at least one estimation error variance, weights, and severity impact to determine the overall "best" (optimal) solution. Each sensor's contribution to a particular solution is assigned a confidence score. In some embodiments, this confidence score is then reflected as an additive metric in the final subset of sensor performances, where each contribution is appropriately weighted and then summed. In some embodiments, the sensor performance with the highest confidence score is selected as the PNT solution.
[0015] Each sensor's performance is assigned a confidence score based on the severity of the impact and the estimated uncertainty. The fused sensor performance is compared to each other and to individual sensors so that each sensor can be assessed for faulty outputs or performance degradation due to interference, jamming, or spoofing (for sensors like GPS / GNSS receivers).
[0016] If it is determined that the performance of the sensor being used is degraded or deviates from other solution sets, it can be ignored, and an alternative sensor performance can be used. The algorithm then selects from the best performance created by the subset of sensors, or the fused sensor performance with the least deviation or degradation. This algorithm can also be used to identify faulty sensors, biased sensors or solutions, and sensors that are externally degraded by interference signals or conditions.
[0017] Figure 1 This is an exemplary system 100 according to the present technology. In some embodiments, system 100 includes a platform 101. Platform 101 may include internal sensors 102A and 102B, at least one external sensor 103, and a controller 104. Although two internal sensors 102A and 102B are illustrated, it should be understood that system 100 may include any number of internal sensors 102. Furthermore, it should be understood that although a single external sensor 103 is illustrated, system 100 may include any number of external sensors 103. Additionally, internal sensors 102 and external sensors 103 may be collectively referred to herein as sensors or multiple sensors. Therefore, the term "sensor" or "multiple sensors" may include internal sensor 102, external sensor 103, or a combination thereof.
[0018] Platform 101 can take any form. In some embodiments, platform 101 is a vehicle. In such embodiments, for example, when platform 101 is a vehicle, platform 101 is mobile. In some embodiments, platform 101 is an aircraft. In some embodiments, platform 101 is a land vehicle. In some embodiments, platform 101 is an aerospace vehicle. In some embodiments, platform 101 is a car, airplane, helicopter, satellite, bus, train, truck, etc.
[0019] In some embodiments, system 100 includes multiple sensors (including internal sensors 102A, 102B and external sensors 103). In some embodiments, internal sensors 102 include accelerometers, gyroscopes, oscillators, wheel tick sensors, etc. In some embodiments, internal sensors 102A, 102B are different sensors, such as an accelerometer and a wheel tick sensor. In some embodiments, internal sensors 102A, 102B include repeating sensors, such as two accelerometers. System 100 may include any number of internal sensors 102.
[0020] In some embodiments, the external sensor 103 is an infrared sensor, an ultrasonic sensor, an optical sensor, or a proximity sensor. In some embodiments, the external sensor 103 is a magnetometer (Hall effect sensor), a barometer, a temperature sensor, a radio frequency sensor, etc. In some embodiments, for example when multiple external sensors 103 are present, the external sensors 103 are different sensors, such as an ultrasonic sensor and a proximity sensor. In some embodiments, the external sensor 103 includes repeating sensors, such as two ultrasonic sensors. The system 100 may include any number of external sensors 103.
[0021] In some embodiments, internal sensors 102A and 102B have a higher weight than external sensor 103. In some embodiments, as described herein, influencing the performance of internal sensors 102A and 102B is more difficult because they are not physically accessible. In contrast, external sensor 103 operates using external stimuli and is therefore physically accessible.
[0022] In some embodiments, system 100 includes controller 104. The controller contains a single Kalman filter (such as...). Figure 2(Shown and described in detail). In some embodiments, the controller 104 is electrically coupled to a plurality of sensors 102A, 102B, 103 via a wired or wireless connection. In operation, the controller 104 can acquire one or more measurements from the plurality of sensors 102A, 102B, 103 attached to the platform 101, determine the estimation error variance of each of the plurality of sensors 102A, 102B, and 103, determine the weight of each of the plurality of sensors 102A, 102B, and 103, and determine the severity of the impact of each of the plurality of sensors 102A, 102B, and 103. In some embodiments, controller 104 is further configured to calculate the individual sensor performance of each of the plurality of sensors 102A, 102B, and 103, based at least on the estimation error variance, weights, and severity effects of each of the plurality of sensors 102A, 102B, and 103. As described herein, controller 104 includes a single Kalman filter. In some embodiments, controller 104 is further configured to generate fused sensor performance by summing the individual sensor performances of all sensors 102A, 102B, and 103 of the plurality of sensors 102A, 102B, and 103 using the single Kalman filter. In some embodiments, controller 104 is further configured to generate a plurality of sub-combined sensor performances using the single Kalman filter. In some embodiments, each sensor in a given sub-combination of sensor performances is characterized by its individual sensor performance, and each sub-combination of sensor performances is determined by summing the weighted individual sensor performances of the individual sensors. Controller 104 may then compare the plurality of sub-combined sensor performances with the fused sensor performance. In some embodiments, each sub-combination sensor performance and the fused sensor performance includes a confidence score based on the estimated error variance, weights, and severity effects of each sensor in each sensor sub-combination and all sensors in the plurality of sensors. The controller 104 can then select the sensor performance with the predetermined confidence score from the plurality of sub-combination sensor performances or the fused sensor performances as the PNT solution. In some embodiments, the predetermined confidence score is the highest confidence score among each sub-combination sensor performance and the fused sensor performance.
[0023] Figure 2 This is an exemplary system 200 according to the present technology. In some embodiments, system 200 includes a platform 201, a plurality of sensors 202A, 202B, 202C, and a Kalman filter 210. It should be understood that the Kalman filter 210 may be part of a controller (such as...). Figure 1 (Controller 104 in the middle).
[0024] In some embodiments, the plurality of sensors 202A, 202B, 202C include internal sensors, external sensors, or combinations thereof, as described herein. In some embodiments, the plurality of sensors 202A, 202B, 202C include accelerometers, gyroscopes, oscillators, wheel speed marking sensors, infrared sensors, ultrasonic sensors, optical sensors, cameras, lidar sensors, proximity sensors, magnetometers, barometers, temperature sensors, radio frequency sensors, or humidity sensors. Although three sensors 202A, 202B, and 202C are illustrated, this is merely exemplary, and the plurality of sensors 202A, 202B, and 202C may include any number of sensors.
[0025] In operation, Kalman filter 210 receives the individual sensor performance of each of a plurality of sensors 202A, 202B, 202C. In some embodiments, a controller (e.g., controller 104) calculates the individual sensor performance of each of the plurality of sensors 202A, 202B, 202C. The individual sensor performance may include at least the estimated error variance, weights, and severity effects of each of the plurality of sensors 202A, 202B, 202C. In some embodiments, Kalman filter 210 generates fused sensor performance 215. Fuded sensor performance 215 includes the individual sensor performance of each of the plurality of sensors 202A, 202B, 202C. In some embodiments, the same Kalman filter 210 also generates sub-combined sensor performances 220A, 220B, 220C. In some embodiments, Kalman filter 210 generates all possible sub-combined sensor performances from m sensors out of a plurality of (n) sensors. For example, for a system 200 including three sensors A, B, and C, such as Figure 2As shown, three sub-combinations are generated: sub-combination 220A of sensor A and sensor B, sub-combination 220B of sensor B and sensor C, and sub-combination 220C of sensor A and sensor C. It should be understood that the Kalman filter 210 can generate all possible sub-combinations of sensor performance including any number of sensors. For example, in a system with five sensors A, B, C, D, and E, the Kalman filter 210 can generate 10 unique sub-combinations (AB, AC, AD, AE, BC, BD, BE, CD, CE, DE) containing two of the five sensors, 10 unique sub-combinations (ABC, ABD, ABE, ACD, ACE, ADE, BCD, BCE, BDE, CDE) containing three of the five sensors, and 5 unique sub-combinations (ABCD, ABCE, ABDE, ACDE, BCDE) containing four of the five sensors. Therefore, the Kalman filter 210 can utilize linear algebra to add additional rows, columns, or rows and columns to the sub-combinations of sensor performance. In this way, a single Kalman filter 210 can generate sub-combinations of all possible sensor performances without using multiple filters or algorithms. In some embodiments, the Kalman filter 210 may compute only a subset of the possible sub-combinations, such as only computing a sub-combination with three sensors out of a plurality of sensors. In some embodiments, the Kalman filter 210 may be another type of filter, such as a particle filter, a least-squares filter, or a Wiener filter.
[0026] Once the fused sensor performance 215 and the sub-combinations 220A, 220B, and 220C of each sensor performance are generated, the system 200 determines which sensor performance among the fused sensor performance 215 and the sub-combinations 220A, 220B, and 220C has a predetermined confidence score, or is closest to a predetermined confidence score. In some embodiments, the predetermined confidence score is the expected confidence score of the system 200 when each of the plurality of sensors 202A, 202B, and 202C is in a fully operational state. In some embodiments, the predetermined confidence score is the highest confidence score from the plurality of sub-combinations of sensor performance 220A, 220B, and 220C or the fused sensor performance 215.
[0027] In some embodiments, the fused sensor performance 215 and multiple sub-combinations of performance 220A, 220B, 220C are generated sequentially. Therefore, in such embodiments, the Kalman filter 210 can switch the selected sensor performance in real time based on which sensor performance in one of the sub-combinations of fused sensor performance 215 or multiple sub-combinations of performance 220A, 220B, 220C is closest to a predetermined confidence score at any given time.
[0028] Figure 3 This is an exemplary method 300 for generating positioning, navigation, or timing solutions according to the present technology.
[0029] In block 305, the system is configured with multiple sensors. This system can be any system described herein, such as system 100, system 200, or system 400. Therefore, the multiple sensors can be any multiple sensors described or illustrated herein, such as multiple sensors 102A, 102B, 103, multiple sensors 202A, 202B, 202C, or multiple sensors 402A, 402B, 402C. In some embodiments, the multiple sensors include one or more internal sensors, one or more external sensors, or a combination thereof.
[0030] Step 310 includes blocks 310A, 310B, and 310C, determining the error variance, weights, and severity impact of each of the plurality of sensors. In some embodiments, the error variance, weights, and severity impact (also referred to herein as “metrics”) are determined by a controller (e.g., controller 104). In some embodiments, blocks 310A, 310B, and 310C occur simultaneously. In some embodiments, blocks 310A, 310B, and 310C occur concurrently. In some embodiments, any one or a combination of blocks 310A, 310B, or 310C may be omitted in step 310.
[0031] In some embodiments, the error variance determined in block 310A is determined within a sigma (1σ) range. In some embodiments, the error variance is estimated based on research or collected data for each individual sensor, and is based on the sensor type. For example, an odometer may have an error variance derived from previous research or data collected from that odometer or other similar odometers.
[0032] In block 310B, a weight is determined for each sensor. In some embodiments, the weights are based on the sensor type of each of the plurality of sensors. In some embodiments, the sensor type is selected from internal sensors or external sensors. In some embodiments, internal sensors have a higher weight than external sensors. For example, an internal sensor may be assigned a weight of 10. In some embodiments, a weight of 10 indicates that the sensor operates without an external signal. In some embodiments, an external sensor may be assigned a weight of 5. In some embodiments, a weight of 5 indicates that the sensor operates in response to an external signal. The purpose of this is to indicate “direct trust” (or sensors that can be trusted with a high degree of confidence with minimal evaluation of other factors) versus “indirect trust” (or sensors that can be trusted based on other factors).
[0033] In box 310C, the severity of the impact on each of the multiple sensors is determined. In some embodiments, the severity of the impact is assigned a numerical value. For example, a value of 1 indicates no impact, a value of 0.75 indicates a minor impact, such as a small inconvenience with usability degradation, a value of 0.5 indicates a significant impact, such as core functionality being affected and undesirable performance degradation, and a value of 0 indicates a severe impact, such as the sensor becoming inoperable. An example of an external factor with a severe impact is radio frequency (RF) interference at a specific RF frequency for a sensor measuring RF. An example of an external factor with a major impact is wheel slippage (mud) for a sensor measuring distance. An example of an external factor with a minor impact is temperature variation for a sensor measuring inertia.
[0034] In block 315, the individual sensor performance of each of the multiple sensors is calculated. Individual sensor performance can be evaluated based at least on the metrics determined in step 310. Therefore, individual sensor performance can be based at least on the error variance of each sensor, the weight of each sensor, and the severity of the performance impact due to one or more external factors. In some embodiments, the evaluations of the metrics are additive and then normalized. Thus, each sensor is assigned a confidence score based on a combination of the estimated error variance, the number of external factors that may affect sensor performance, and the impact of each normalization. A higher confidence score indicates a greater likelihood that the sensor is trustworthy (i.e., the confidence score is closer to a predetermined confidence score).
[0035] External influences can negate the use of sensors. For example, if RF interference is present and detected by the system, sensors using RF technology in that frequency band will be inoperable. Therefore, the observed impact of external factors on the performance of a single sensor is a multiplier of the final confidence score. In some embodiments, the performance of a single sensor is calculated using Equation 1.
[0036]
[0037] Equation 1
[0038] As shown in Equation 1, E is the variance of the estimation error. p This refers to the number of external factors that may affect sensor performance. p O is the severity of the performance impact caused by one or more external factors. T is the type of sensor, and O is the observed impact on sensor performance caused by one or more external factors. In some embodiments, the confidence scoring process is performed before Kalman filtering, but is represented as a floating-point number in the sub-combination of sensor performance.
[0039] In block 320, the fused sensor performance is generated. In some embodiments, a Kalman filter generates the fused sensor performance. In some embodiments, the fused sensor performance is generated by a controller (e.g., controller 104). In some embodiments, the fused sensor performance (also referred to as the fused solution) is generated using Equation 2.
[0040]
[0041] Equation 2
[0042] As shown in Equation 2, N is the number of sensors.
[0043] In box 330, the scores of multiple sub-combinations are compared with the performance of the fused sensor.
[0044] In box 335, a sensor performance is selected from multiple sub-combined sensor performances or fused sensor performances. The fused sensor performance can be selected when its confidence score is closest to a predetermined confidence score (at which all sensors operate without external influence (e.g., when the severity influence of all sensors is 1)) or when one or more sensors operate with only minor influence. If the observed influence of a single sensor performance (wheel slippage, RF interference, etc.) causes a sensor to be inoperable or severely affected, the fused sensor performance may not be used. Instead, the system can select a sub-combined sensor performance that is closer to the predetermined confidence score. In some embodiments, the system will select the sub-combined sensor performance with the highest confidence score.
[0045] Figure 4 This is a block diagram of an exemplary system 400 according to the present technology. In some embodiments, system 400 includes a plurality of sensors 402A, 402B, 402C. In some embodiments, system 400 generates individual sensor performance for each of the plurality of sensors 402A, 402B, 402C.
[0046] In blocks 405A-1, 405A-2, and 405A-3, the error variance of each of the multiple sensors 402A, 402B, and 402C is added to the performance of a single sensor. In blocks 405B-1, 405B-2, and 405B-3, the severity effect is subtracted from the performance of the single sensor of each of the multiple sensors 402A, 402B, and 402C. In some embodiments, blocks 405A-1, 405A-2, and 405A-3 occur concurrently with blocks 405B-1, 405B-2, and 405B-3. In some embodiments, blocks 405A-1, 405A-2, and 405A-3 occur concurrently with blocks 405B-1, 405B-2, and 405B-3.
[0047] In boxes 410-1, 410-B, and 410-C, each of the multiple sensors 402A, 402B, and 402C is weighted according to the severity of the impact.
[0048] In some embodiments, the performance of a single sensor is generated continuously. In such embodiments, as shown in blocks 415-1, 415-2, and 415-3, one or more external factors (as described herein) that significantly affect the impact of the multiple sensors 402A, 402B, 402C are monitored. Monitoring of the one or more external sensors can be continuous, so the severity of the impact is measured in real time. Therefore, the performance of a single sensor can be updated in real time based on the monitored external factors.
[0049] In boxes 420-1, 420-2, and 420-3, individual sensor performance is output for each sensor. Individual sensor performance can be updated in real time based on monitored external factors.
[0050] Example
[0051] The following provides conceptual examples of integrating solutions and scoring metrics in different environments.
[0052] In this example, the fused sensor performance consists of measurements from three sensors: an inertial measurement unit (IMU), a GPS receiver, and an odometer. The IMU provides inertial acceleration and gyroscope measurements, the GPS receiver provides position and velocity information, and the odometer provides velocity measurements.
[0053] In this example, system 400 is used under three specific conditions: normal GPS signal conditions, interfered GPS conditions with minor signal interference, and GPS denial conditions with severe signal interference.
[0054] Each sensor either provides an estimated one-sigma variance or estimates that one-sigma variance internally. In this example, the variances calculated for the IMU, GPS receiver, and odometer in each environment are 5.0, 1.0, and 7.0, respectively. The IMU and odometer are internal sensors with a sensor weight of 10, while the GPS receiver is an external sensor with a sensor weight of 5. Furthermore, the odometer and IMU operate normally without any performance impact from external factors. The GPS RF environment affects performance and causes varying degrees of severity. For normal GPS conditions, the value is 1. For jammed GPS environments, the value is 0.5. For GPS-denied environments, the value is 0.
[0055] Tactical-level IMUs are affected by external factors with associated severity. In this example, the IMU is affected by the following: temperature: 0.75 (minor); high vibration: 0.75 (minor); magnetic fluctuation: 0.75 (minor).
[0056] GNSS receivers are also affected by external factors with associated severity. In this example, the GNSS receiver is affected by the following: Electromagnetic interference: 0.5 (significant); RF interference: 0 (inoperable).
[0057] Wheel speed odometers are affected by external factors with associated severity. Here, wheel speed odometers are affected by the following: high vibration: 0.75 (minor); wheel slippage: 0.5 (major); electromagnetic interference: 0.75 (minor). This list is by no means exhaustive, but merely an example of potential system variations.
[0058] The individual sensor performance is calculated for each sensor as shown below. In the example below, Equation 1 is used to calculate the performance of each individual sensor.
[0059] In a normal GPS environment, the performance of each individual sensor is calculated as shown in Equations 3, 4, and 5.
[0060] IMU sensor performance = (1)*(5–10 +0.75) = -5.75
[0061] Equation 3
[0062] Odometer sensor performance = (1)*(7–10 +.67) = -3.67
[0063] Equation 4
[0064] GPS receiver sensor performance = (1)*(1–5 +.25) = -4.25
[0065] Equation 5
[0066] In a jammed GPS environment, the performance of each individual sensor is calculated as shown in Equations 6, 7 and 8.
[0067] IMU sensor performance = (1)*(5–10 + 0.75) = -5.75
[0068] Equation 6
[0069] Odometer sensor performance = (1)*(7–10 + 0.67) = -3.67
[0070] Equation 7
[0071] GPS receiver sensor performance = (0.5)*(1–5 + 0.25) = -1.875
[0072] Equation 8
[0073] Finally, in a GPS denied environment, the performance of each individual sensor is calculated as shown in Equations 9, 10, and 11.
[0074] IMU sensor performance = (1)*(5–10 + 0.75) = -5.75
[0075] Equation 9
[0076] Odometer sensor performance = (1)*(7–10 + 0.667) = -3.67
[0077] Equation 10
[0078] GPS receiver sensor performance = (0)*(1–5 + 0.25) = 0
[0079] Equation 11
[0080] Then, multiple sub-combinations of sensor performance are generated for each case, as shown below. The sub-combinations of sensor performance for a single sensor are evaluated based on the individual sensor performance score. The sub-combinations of sensor performance are as follows: IMU and GPS, IMU and odometer, and GPS and odometer.
[0081] In a normal GPS environment, the performance of each sub-sensor combination is calculated as shown in Equations 12, 13 and 14.
[0082] IMU and GPS receiver performance = (-5.75 + -4.25) = -10
[0083] Equation 12
[0084] IMU and odometer performance = (-5.75 + -3.67) = -9.42
[0085] Equation 13
[0086] GPS receiver and odometer performance = (-4.25 + -3.67) = -7.92
[0087] Equation 14
[0088] Then, Equation 2 is used to calculate the fused sensor performance, which is the sum of the performance metrics of all individual sensors, as shown in Equation 15.
[0089] Fusion performance = (-5.75 + -4.25 + -3.67) = -13.67
[0090] Equation 15
[0091] Because the fusion sensor performance has the lowest score, it has the highest confidence score, and in this example, the fusion sensor performance is selected as the solution.
[0092] In a jammed GPS environment, the performance of each sub-sensor combination is calculated as shown in Equations 16, 17 and 18.
[0093] IMU and GPS receiver performance = (-5.75 + -1.875) = -7.625
[0094] Equation 16
[0095] IMU and odometer performance = (-5.75 + -3.67) = -9.42
[0096] Equation 17
[0097] GPS receiver and odometer performance = (-1.875 + -3.67) = -5.545
[0098] Equation 18
[0099] Then, Equation 2 is used to calculate the fused sensor performance, which is the sum of the performance metrics of all individual sensors, as shown in Equation 19.
[0100] Fusion performance = (-5.75 + -1.875 + -3.67) = -11.295
[0101] Equation 19
[0102] Here, the fused sensor performance still provides the lowest score, thus providing the highest confidence score, so the fused sensor performance is selected as the optimal solution.
[0103] In a GPS-denied environment, the performance of each sub-sensor combination is calculated as shown in Equations 20, 21, and 22.
[0104] IMU and GPS receiver fusion performance = (-5.75 + 0) = -5.75
[0105] Equation 20
[0106] IMU and odometer fusion performance = (-5.75 + -3.67) = -9.42
[0107] Equation 21
[0108] GPS receiver and odometer fusion performance = (-0 + -3.67) = -3.67
[0109] Equation 22
[0110] Then, Equation 2 is used to calculate the fused sensor performance, which is the sum of the performance metrics of all individual sensors, as shown in Equation 23.
[0111] Fusion sensor performance = (-5.75 + 0 + -3.67) = -9.42
[0112] Equation 23
[0113] Here, the fusion of sensor performance includes sensors (GPS receivers) that are inoperable due to severe external performance factors. The sub-combination of IMU and odometer has the lowest score (i.e., the highest confidence score). Therefore, the sub-combination of IMU and odometer is chosen as the solution here.
[0114] The foregoing description of the illustrated examples of the present invention, including the content described in the abstract, is not intended to be exhaustive or to limit the invention to the precise forms disclosed. While specific examples of the invention have been described herein for illustrative purposes, various modifications can be made within the scope of the invention, as will be recognized by those skilled in the art.
[0115] These modifications can be made to the invention based on the above detailed description. The terminology used in the following claims should not be construed as limiting the invention to the specific examples disclosed in the specification. Rather, the scope of the invention is fully defined by the following claims, which should be interpreted according to the established principles of their application.
Claims
1. A method for providing a positioning, navigation, or timing (PNT) solution, the method comprising: Acquire one or more measurements from multiple sensors attached to, remote from, or in combination thereof on the platform; Determine the variance of the estimation error for each of the plurality of sensors; Determine the weight of each of the plurality of sensors; Determine the severity of the impact of each of the plurality of sensors; The individual sensor performance of each of the plurality of sensors is calculated based at least on the estimated error variance, the weights, and the severity impact of each of the plurality of sensors; Using filters, fused sensor performance is generated by adding the individual sensor performance of each of the plurality of sensors; The filter is used to generate multiple sub-combinations of sensor performance, wherein each sensor in a given sub-combination of sensor performance is characterized by its individual sensor performance, and wherein each sub-combination of sensor performance is determined by weighted summing of the individual sensor performances of the individual sensors in the given sub-combination. The performance of the plurality of sub-combined sensors is compared with the performance of the fused sensor, wherein the performance of each sub-combined sensor and the performance of the fused sensor have a confidence score based on the estimated error variance, the weight, and the severity impact of each of the plurality of sensors; and Select sensor performance with a predetermined confidence score from the plurality of sub-combined sensor performances or the fused sensor performances.
2. The method according to claim 1, wherein, The predetermined confidence score is the highest confidence score derived from the performance of the plurality of sub-combined sensors or the fused sensor performance.
3. The method according to claim 1, wherein, The fused sensor performance and the multiple sub-combination performance are generated sequentially, and wherein the filter is a single Kalman filter.
4. The method according to claim 3, wherein, The method further includes: switching sensor performance in real time based on which sensor performance has the highest confidence score among the fused sensor performance or the performance of one of the multiple sub-combinations.
5. The method according to claim 1, wherein, The multiple sensors include accelerometers, gyroscopes, oscillators, wheel speed marking sensors, infrared sensors, ultrasonic sensors, optical sensors, cameras, lidar sensors, proximity sensors, magnetometers, barometers, temperature sensors, radio frequency sensors, or humidity sensors.
6. The method according to claim 1, wherein, The severe impacts include: One or more external factors; and The severity of the impact caused by the one or more external factors mentioned above.
7. The method according to claim 6, wherein, The severity of the impact is adjusted based on observations of the one or more external factors.
8. The method according to claim 6, wherein, The one or more external factors are selected from radio frequency jamming, spoofing, visible degradation, sensor damage, temperature, wheel slippage, or a combination thereof.
9. The method according to claim 6, wherein, The performance of a single sensor is determined by the following formula: Where E is the variance of the estimation error; N p The quantity of the one or more external factors; S p The severity of the impact caused by the one or more external factors; T represents the type of sensor; and O represents the observed effect caused by the one or more external factors mentioned above.
10. The method according to claim 9, wherein, The type of sensor is selected from internal or external sensors.
11. The method according to claim 10, wherein, The internal sensor has a higher weight than the external sensor.
12. A system for generating positioning, navigation, or timing (PNT) solutions, the system comprising: Multiple sensors are configured to perform one or more measurements; as well as The controller is configured to: Receive one or more measurement values from the plurality of sensors; Determine the variance of the estimation error for each of the plurality of sensors; Determine the weight of each of the plurality of sensors; Determine the severity of the impact of each of the plurality of sensors; The individual sensor performance of each of the plurality of sensors is calculated based at least on the estimated error variance, the weights, and the severity impact of each of the plurality of sensors; Using filters, fused sensor performance is generated by summing the individual sensor performance of each of the plurality of sensors; The filter is used to generate a plurality of sub-combinations of sensor performance, wherein each sensor in the sub-combination of sensor performance is characterized by its individual sensor performance, and wherein each sub-combination of sensor performance is determined by adding the weighted individual sensor performances of the individual sensors belonging to the sub-combination of each sensor performance belonging to a given sub-combination. The performance of the plurality of sub-combined sensors is compared with the performance of the fused sensor, wherein the performance of each sub-combined sensor and the performance of the fused sensor have a confidence score based on the estimated error variance, the weight, and the severity impact of each of the plurality of sensors; and Select sensor performance with a predetermined confidence score from the plurality of sub-combined sensor performances or the fused sensor performances.
13. The system according to claim 12, wherein, The controller is further configured to continuously generate the fused sensor performance score and multiple sub-combination scores, wherein the filter is a single Kalman filter.
14. The system according to claim 13, wherein, The controller is further configured to switch sensor performance in real time based on which sensor performance has the highest confidence score among the fused sensor performance or the performance of one of the multiple sub-combinations.
15. The system according to claim 12, wherein, The multiple sensors include accelerometers, gyroscopes, oscillators, wheel speed marking sensors, infrared sensors, ultrasonic sensors, optical sensors, cameras, lidar sensors, proximity sensors, magnetometers, barometers, temperature sensors, radio frequency sensors, or humidity sensors.
16. The system according to claim 12, wherein, The severe impacts include: One or more external factors; and The severity of the impact caused by the one or more external factors mentioned above.
17. The system according to claim 16, wherein, The severity of the impact is adjusted based on observations of the one or more external factors.
18. The system according to claim 16, wherein, The one or more external factors are selected from radio frequency jamming, spoofing, visible degradation, sensor damage, temperature, wheel slippage, vibration, magnetic distortion, electromagnetic interference, or a combination thereof.
19. The system according to claim 16, wherein, The performance of a single sensor is determined by the following formula: Where E is the variance of the estimation error; N p The number of one or more external factors that affect the performance of the individual sensor; S p The severity of the impact due to one or more external factors that may affect the performance of a single sensor; T represents the type of sensor; and O represents the observed effect caused by the one or more external factors mentioned above.
20. The system according to claim 19, wherein, The type of sensor is selected from internal or external sensors.