A multi-robot system fault detection method against external disturbance influences

Abstract

The application relates to a kind of multi-robot system fault detection methods for external interference influence, comprising: the position sensor is equipped on each robot in multi-robot system, the relative distance information of itself and neighbor robot is measured in real time by position sensor;According to the relative distance information measured by each robot, the fault detection reference model of corresponding robot is respectively constructed, and the parameter design of the improved intermediate observer corresponding to each robot is completed;Based on the intermediate observer of determined parameter, the estimated value of intermediate observer is obtained, and suitable residual generation algorithm is selected, residual generation is carried out according to the estimated value of intermediate observer;Corresponding residual evaluation rule is used to carry out fault detection.Compared with prior art, the application has the advantages of simple equipment, high detection precision, high robustness and the like.

Description

Technical Field

[0001] The present invention relates to the field of robot system fault detection, and in particular to a multi-robot system fault detection method against external interference. Background Art

[0002] A robot can be considered an independent system that perceives its environment, acquires information, interacts with it, and completes assigned tasks using stored prior knowledge or continuously updated judgments and strategies. Through training, individual robots can gradually improve and enhance their capabilities through interaction with their environment to complete tasks. However, in more complex environments, the number of robots is often not limited to a single one, and the capabilities of a single robot alone are often insufficient to perform tasks in complex environments. A natural solution is to increase the number of robots, combining multiple trained robots to form a multi-robot system. This overcomes the problem of a single robot's limited capabilities in complex environments and allows it to adapt to even more complex environments. In this research field, with the advancement of artificial intelligence, particularly deep learning, research and applications of multi-robot systems have increased, such as in power system diagnosis, drone formations, multi-manipulator collaborative assembly, and satellite formations.

[0003] A multi-robot system consists of multiple robots interacting and collaborating. This means that if one robot fails, the entire system may crash. This characteristic of multi-robot systems has led to higher demands for their safety and reliability, and fault detection in multi-robot systems has gradually attracted the attention of scholars. With the deepening of related research, fault detection methods for multi-robot systems are also increasing. These methods can generally be divided into the following three categories: model-based fault detection methods, data-driven fault detection methods, and other fault detection methods.

[0004] (1) Model-based fault detection method

[0005] Model-based fault detection methods are developed for systems with known models. They primarily consist of observer-based fault detection methods, including those based on unknown input observers, sliding mode observers, and Kalman filters. The core concept is to establish a suitable observer to estimate a characteristic parameter of the system and then compare the estimated value with the actual value. This comparison is typically performed by performing a direct difference, generating a detection metric known as a residual error. This metric is then used to determine whether the system has experienced a fault.

[0006] (2) Data-driven fault detection method

[0007] Data-driven fault detection methods are developed for systems with unknown models. Due to the difficulty of obtaining relevant dynamic models in complex process, or the tediousness, time-consuming and inaccuracy of modeling process, data-driven fault detection methods are favored by researchers. The traditional data-driven methods, principal component analysis (PCA) and partial least squares (PLS), are the most widely used, and the more novel subspace aided data-driven method is also used.

[0008] (3) Other fault detection methods

[0009] With the rapid development of deep learning and artificial intelligence, fault detection in multi-robot systems has made significant progress. Traditional multi-robot systems often rely on rules and experience for fault diagnosis, while now, combined with deep learning technology, multi-robot systems can more intelligently detect faults in dynamic and complex environments. For example, some methods combine traditional fault diagnosis rules with the learning ability of deep learning. In multi-robot systems, deep learning models can learn and identify fault patterns from a large amount of real-time data, thereby improving the accuracy and speed of fault detection.

[0010] However, the existing multi-robot system fault detection algorithm still has the defects of relying on absolute information sensors, high overall complexity, being easily disturbed, and poor accuracy. SUMMARY

[0011] The purpose of the present application is to overcome the defects of the prior art, such as relying on absolute information sensors, high overall complexity, being easily disturbed, and poor accuracy, and to provide a multi-robot system fault detection method for external interference.

[0012] The purpose of the present application can be achieved by the following technical solutions:

[0013] A multi-robot system fault detection method for external interference, comprising the following steps:

[0014] Position sensors are installed on each robot in the multi-robot system, and the relative distance information of the robot itself and its neighbor robots is measured in real time by the position sensors;

[0015] According to the relative distance information measured by each robot, a fault detection reference model corresponding to each robot is constructed, and the parameter design of the improved intermediate observer corresponding to each robot is completed;

[0016] Based on the determined parameters of the intermediate observer, the estimated value of the intermediate observer is obtained, and a suitable residual generation algorithm is selected to generate residuals according to the estimated value of the intermediate observer; the corresponding residual evaluation rule is used for fault detection.

[0017] Furthermore, the expression of the fault detection reference model of robot i is:

[0018]

[0019] Where A, B, B d 、B f , C is the system matrix, its value is determined by the physical mechanical structure of robot i itself, ΔA i For model heterogeneity between robots, ΔA i =N i F i E i , N i and E i is a known matrix, F i Satisfy F i T F i ≤I, I is the identity matrix, x i (t),u i (t) and Z i (t) represents the motion state, control signal and relative output information of robot i at the current moment, d i (k) and f i (k) are interference signal and fault signal respectively, is the set of all neighbor robots of robot i, x j (t) is the motion state of the neighbor robot j at the current moment, is the derivative of the motion state of robot i at the current moment.

[0020] Based on this, the reference model of the robot i can be transformed into the following form:

[0021]

[0022] Where, The matrix R is a reversible matrix and can be reasonably selected in combination with the following design.

[0023] Furthermore, the expression of the improved intermediate observer is:

[0024]

[0025] In the formula, K1, K2, H1, H2 are the parameter variables to be designed. is the estimated value of the motion state of robot i at the current moment, ξ i (t) represents the intermediate value of robot i at the current moment, expressed as ξ i (t) = r i (t)-θR T x i(t), θ is a positive scalar, and is ξ i is the estimate of (t), ζ i (t) is the output error feedback of robot i at current time, expressed as is the derivative term of error feedback of robot i at current time, is the derivative term of the estimate of motion state of robot i at current time, is the estimate of residual signal of robot i at current time, which is a mixed signal containing heterogeneous information and fault information, i.e. is the heterogeneous information of robot i at current time.

[0026] Further, the residual generation algorithm is to isolate the fault signal by setting the orthogonal complement space, and the corresponding calculation expression is:

[0027]

[0028] wherein, is the estimate of is the residual signal obtained by isolating the fault signal, is the orthogonal complement space of robot i.

[0029] Further, the residual generation algorithm is to multiply vR on both sides of to obtain the estimate of the balanced intermediate observer, and the corresponding calculation expression is:

[0030]

[0031] wherein, is the residual signal obtained by balancing v is the to-be-designed adjustment vector.

[0032] Further, the residual generation algorithm also sets a performance index in the processing process to assist in the design of the vector v, so as to balance between fault sensitivity and robustness to heterogeneous information, and the calculation expression of the performance index is:

[0033]

[0034] wherein, J is the performance index.

[0035] Further, the expression of the residual evaluation rule is:

[0036] If If

[0037] wherein, is the residual result of robot i at the current moment generated by the residual generation algorithm, J th is the threshold function.

[0038] Furthermore, the threshold function J th The choice is based on minimizing the false alarm rate while maximizing the sensitivity to failures.

[0039] Furthermore, the position sensor equipped on each robot is a visual sensor, a radar or a laser radar.

[0040] Furthermore, the multi-robot system is a heterogeneous multi-robot system or a homogeneous multi-robot system.

[0041] Compared with the prior art, the present invention has the following advantages:

[0042] (1) The present invention provides a robust fault detection method for a heterogeneous multi-robot system affected by external interference, which can achieve accurate fault detection by relying only on the relative measurement information of onboard sensors. It also makes full use of the homogeneous and heterogeneous information among multiple robots to achieve high-precision and robust fault detection.

[0043] (2) The fault detection method of the present invention eliminates the influence of external interference on fault detection by designing an improved intermediate observer, thereby ensuring the reliability of the fault detection result.

[0044] (3) The fault detection algorithm provided by the present invention designs a special residual generation algorithm to fully reduce the interference of heterogeneous information among multiple robots on fault signal identification, thereby ensuring the accuracy of the fault detection results.

[0045] (4) The present invention implements a two-step heterogeneous multi-robot fault detection strategy, provides a new solution for heterogeneous multi-robot fault detection, and provides a new approach to the problems caused by different models among multiple robots. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A schematic flow chart of a multi-robot system fault detection method for external interference provided in an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram of an actual application scenario provided in an embodiment of the present invention, including five wheeled robots traveling in formation, where robot 1 has a different structure from the other robots;

[0048] Figure 3 This is a schematic diagram of the change in formation when a robot fails during the marching of a multi-robot formation provided in an embodiment of the present invention;

[0049] Figure 4A schematic diagram of a change in the connection relationship of a multi-robot formation after the fault detection algorithm of the present invention takes effect is provided in an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of the re-formation of the remaining robots after a faulty robot is located and isolated according to an embodiment of the present invention;

[0051] Figure 6 This is a diagram showing the estimation effect of an improved intermediate observer provided in an embodiment of the present invention and an original method under strong interference;

[0052] Figure 7 A comparison diagram of the effects of an existing method of the fault detection method according to the present invention provided in an embodiment of the present invention;

[0053] Figure 8 This is a diagram showing the effect of the residual generation algorithm 1 described in the present invention failing under special circumstances, provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0055] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0056] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0057] Example 1

[0058] like Figure 1 As shown, this embodiment provides a multi-robot system fault detection method for external interference, including the following steps:

[0059] S1: Each robot in the multi-robot system is equipped with a position sensor, which measures the relative distance between itself and its neighboring robots in real time.

[0060] S2: Based on the relative distance information obtained by each robot, a fault detection reference model for each robot is constructed, and the parameters of the improved intermediate observer corresponding to each robot are designed.

[0061] S3: Based on the intermediate observer with determined parameters, obtain the value estimated by the intermediate observer, select an appropriate residual generation algorithm, and generate residuals according to the value estimated by the intermediate observer; use the corresponding residual evaluation rule to perform fault detection.

[0062] In step S1, the position sensor equipped on each robot is a visual sensor, radar or lidar.

[0063] In a multi-robot system, each robot is equipped with sensors such as vision, radar, or lidar to measure the relative distance to its neighbors in real time, rather than relying on traditional absolute positioning methods like GPS. Vision sensors enable robots to identify their surroundings and estimate their relative positions, while radar or lidar provides highly accurate distance measurements. These sensors enable robots to autonomously locate themselves even in environments without GPS signals, making them particularly useful indoors or in complex terrain. Collaboration and information sharing between robots enhances system robustness, ensuring that the overall system remains operational even if some robots fail.

[0064] In step S2, the fault detection reference model describes the current motion state of the robot, and its construction relies only on relative position information. The relative information is combined with the consensus control protocol to couple the states of the robot and its neighbors. This requires the mutual detection of neighbor faults in the multi-robot system and further processing of heterogeneous system models. The heterogeneity of the model is usually manifested as order differences and parameter changes between agents. In order to standardize agents with different orders, it can be achieved by adding additional patterns to low-order agents. For example, the state space representation of the low-order robot can be adjusted by adding the identity matrix to the matrix A. Parameter changes in other system matrices can be standardized by appropriate parameter transformations.

[0065] Therefore, the fault detection reference model of robot i is fully described by the following mathematical model:

[0066]

[0067] Where A, B, B d 、B f , C is the system matrix, its value is determined by the physical mechanical structure of robot i itself, ΔA i For model heterogeneity between robots, ΔA i =N i F i E i , Ni and E i is a known matrix, F i Satisfy F i T F i ≤I, I is the identity matrix, x i (t),u i (t) and Z i (t) represents the motion state, control signal and relative output information of robot i at the current moment, d i (k) and f i (k) are interference signals and fault signals respectively. When the robot fails, it satisfies f i (t)≠0, is the set of all neighbor robots of robot i, x j (t) is the motion state of the neighbor robot j at the current moment, is the derivative of the motion state of robot i at the current moment.

[0068] Based on this, the reference model of the robot i can be transformed into the following form:

[0069]

[0070] Where, The matrix R is a reversible matrix and can be reasonably selected in combination with the following design.

[0071] The expression of the improved intermediate observer is:

[0072]

[0073] In the formula, K1, K2, H1, H2 are the parameter variables to be designed. is the estimated value of the motion state of robot i at the current moment, ξ i (t) represents the intermediate value of robot i at the current moment, expressed as ξ i (t) = r i (t)-θR T x i (t), θ is a positive scalar, and for ξ i The estimated value of (t), ζ i (t) is the output error feedback between robot i and its neighbors at the current moment, expressed as is the derivative of the error feedback of robot i at the current moment, is the derivative of the estimated value of the motion state of robot i at the current moment, is the estimated value of the residual signal of robot i at the current moment, which contains a mixed signal of heterogeneous information and fault information, that is, is the heterogeneous information of robot i at the current moment.

[0074] The innovation of the proposed intermediate observer is reflected in two key aspects. First, compared with the traditional state observer, the error feedback ζ i (t) incorporates state information from all neighboring robots. This introduces homogeneous information, or hardware redundant information, into the observer, thereby enhancing its potential for performance optimization. Secondly, the derivative term of the error feedback is added. This provides an opportunity to improve the dynamic characteristics of the observer.

[0075] In step S3, the main task of the fault detection algorithm is divided into two steps: residual generation and residual evaluation. By running the intermediate observer in the previous step, the variable of heterogeneous information and fault signal mixture can be obtained. Right now

[0076]

[0077] The goal of the residual generation algorithm is to make the residual signal contain only fault information, or at least make the fault signal more prominent in the residual, while minimizing the impact of other interference signals, thereby improving the sensitivity and accuracy of fault detection.

[0078] Heterogeneous parameter moment N i It is usually rank-deficient, which means it has an orthogonal complement space Through the signal The residual generation algorithm 1 can be obtained by processing:

[0079]

[0080] Where, For The residual signal obtained by isolating the fault signal, is the orthogonal complementary space of robot i.

[0081] This result shows that the fault signal can be successfully isolated. This method significantly reduces the impact of heterogeneous information, thereby improving the accuracy of fault detection. However, this may also bring the risk of losing some fault information, which in turn leads to missed detection. For example, when the matrix The fault signal may be inadvertently eliminated when

[0082] To overcome this limitation, a balanced approach is adopted, which aims to optimize the sensitivity to faults and enhance the robustness to heterogeneous information to achieve higher detection performance. Specifically, multiplying both sides of Equation (4) by vR yields the residual generation algorithm 2:

[0083]

[0084] Where, For balance The residual signal obtained after , v is the adjustment vector to be designed.

[0085] In order to strike a balance between fault sensitivity and robustness to heterogeneous information, the following commonly used performance indicators are proposed:

[0086]

[0087] Where J is the performance index, which can assist in the design of the vector v.

[0088] The residual generation methods discussed above are suitable for different application scenarios. Residual generation methods based on orthogonal complement space produce the best results when the fault signal is completely independent of heterogeneous information. In contrast, residual generation methods based on performance indicators have greater versatility, although their effectiveness in fault detection may be slightly reduced. Because fault signals and heterogeneous information are often unknown in real systems, a series of tests can be performed to determine the most appropriate residual generation method.

[0089] After generating the residual signal, a reasonable residual evaluation strategy must be used to verify the fault. The commonly used residual evaluation logic is as follows:

[0090] like

[0091] Where, is the residual result of robot i at the current moment generated by the residual generation algorithm, J th is the threshold function. Threshold J th The choice of should aim to minimize the false alarm rate while maximizing the sensitivity to failures.

[0092] This method is suitable for heterogeneous multi-robot fault detection application scenarios, and can also be applied to homogeneous multi-robot fault detection scenarios. It can realize robust multi-robot fault detection when only relative measurement information is available. It has the advantages of simple equipment, high detection accuracy and high robustness.

[0093] Specific implementation process:

[0094] The present invention is described below using five wheeled robots traveling in formation as an example, and specifically includes the following steps:

[0095] Step 1: The robot uses onboard sensors to measure the position information of itself and neighboring robots in real time. Figure 2As shown, robot 1 is equipped with an on-board radar, and robots 2-5 are equipped with on-board cameras. The red solid line indicates that the two robots are neighbors and can measure each other's position information in real time.

[0096] Step 2: Start all robots and the fault detection algorithm of the present invention, and build a robot fault detection reference model with reference to formula (1). In this embodiment, the robot fault detection reference model is as follows:

[0097]

[0098] where x i (t)=[p i (t)v i (t)] T , p i (t), v i (t) represents the position and velocity information of the robot respectively, and the rest of the symbols and variable definitions are exactly the same as in (1).

[0099] Step 3: Construct a residual generator as shown in formula (3) and determine the residual generation coefficients K1, K2, H1, and H2. In this embodiment, the residual generation coefficients K1, K2, H1, and H2 are selected as:

[0100]

[0101] In this embodiment, during the period of 0-20 seconds, a fault occurs in the robot car 1 at 10 seconds, which is manifested as a step fault signal. The detection effect of the improved intermediate observer and the original intermediate observer is as follows: Figure 5 The numerical calculation results show that Figure 6 The root mean square error between the estimated value and the actual value of the proposed method is 0.17958, while the root mean square error of the original observer is 0.23056. The above results show that the proposed improved intermediate observer has stronger anti-interference robustness than the original observer.

[0102] Step 4: Design the residual generation algorithm 1 according to formula (5), and get the matrix The value of is chosen as:

[0103]

[0104] Step 5: Design the residual generation algorithm 2 by referring to formula (6), and the value of variable v can be selected as: v = [1.35670.1251] T . Design the threshold function J according to the appropriate method th For comparison, when r(t)≤J th When r(t)>Jth When , it can be determined that there is a fault in the robot itself or in the neighboring robots. Because the threshold function values ​​of different methods are different, in order to better compare the effects of each method, the threshold function value is unified to 1, and the other values ​​are scaled proportionally.

[0105] Step 6: During the multi-robot formation operation, the above residual generation and evaluation algorithm is run in real time. Figure 7 , we can conclude that at 8 seconds, faults begin to occur, causing the residuals of all methods to exceed the threshold, demonstrating the fault detection performance of each method. Between 19 and 24 seconds, due to the influence of heterogeneous parameters, the residuals of the original intermediate estimator and the Kalman filter-based methods intermittently drop below the threshold, indicating an increased likelihood of missed detection. In contrast, the two residual generation methods proposed in this paper, especially the method based on residual generation algorithm 1, demonstrate strong robustness to parameter heterogeneity, and the overall waveform shows relatively stable characteristics.

[0106] Step 7: It should be noted that, as discussed previously, the method based on the residual generation algorithm 1 does not always provide detection performance that is superior to other methods. It has the risk of losing key information, which may lead to missed detections.

[0107] For example, when

[0108]

[0109] The results of the four methods are as follows Figure 8 As shown in the figure, while the other three methods effectively detected faults, the method based on residual generation algorithm 1 failed to detect the fault. These results indicate that although the proposed methods based on residual generation algorithm 1 and residual generation algorithm 2 outperform traditional methods in fault detection performance, their applicability varies. In practical applications, the most appropriate method must be selected based on thorough analysis or testing.

[0110] like Figure 3-5 FIG. 1 is a diagram illustrating the changing process of a multi-robot formation after adopting the fault detection method of the present invention.

[0111] The present invention provides a multi-robot rapid fault detection method based on local measurement, which has the following outstanding innovations compared to existing methods: First, the robust fault detection method for heterogeneous multi-robot systems affected by external interference relies solely on the relative measurement information of onboard sensors to achieve accurate fault detection, and fully utilizes the homogeneous and heterogeneous information between multiple robots to achieve high-precision and robust fault detection. Second, the fault detection method of the present invention eliminates the influence of external interference on fault detection by designing an improved intermediate observer, thereby ensuring the reliability of the fault detection results. Then, the fault detection algorithm of the present invention designs a special residual generation algorithm to fully reduce the interference of heterogeneous information between multiple robots on fault signal identification, thereby ensuring the accuracy of the fault detection results. Finally, the two-step heterogeneous multi-robot fault detection strategy provides a new solution for heterogeneous multi-robot fault detection and provides a new approach to problems caused by different models between multiple robots.

[0112] In summary, the present invention achieves robust fault detection and localization of heterogeneous multi-robots under interference through multiple innovations such as a relative measurement-based method, an improved intermediate observer, and the application of a novel residual generation algorithm. It has broad application prospects and important social significance.

[0113] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A multi-robot system fault detection method for external interference, characterized in that: The following steps are involved: Each robot in the multi-robot system is equipped with a position sensor, which measures the relative distance between itself and its neighboring robots in real time. Based on the relative distance information obtained by each robot, a fault detection reference model for the corresponding robot is constructed, and the parameters of the improved intermediate observer corresponding to each robot are designed; Based on the intermediate observer with determined parameters, the estimated value of the intermediate observer is obtained, and a suitable residual generation algorithm is selected to generate residuals according to the estimated value of the intermediate observer; and the corresponding residual evaluation rule is used to perform fault detection; robot The expression of the fault detection reference model is: Where, 、 、 、 、 is the system matrix, whose value is determined by the robot Its own physical and mechanical structure determines For model heterogeneity between robots, , and is a known matrix, satisfy , is the identity matrix, 、 and Represents the current moment robot The motion state, control signal and relative output information, and are interference signal and fault signal respectively, For robots The set of all neighbor robots of Neighbor robot at the current moment state of motion, For the current moment robot The derivative of the motion state; Based on this, the above robot The fault detection reference model is transformed into the following form: Where, ,matrix is a reversible matrix, For the current moment robot heterogeneous information; The expression of the improved intermediate observer is: Where, 、 、 、 is the parameter variable to be designed, For the current moment robot The estimated value of the motion state, Represents the robot at the current moment The middle value of , is a positive scalar, and for The estimated value of For the current moment robot The output error feedback of its neighbors is expressed as , For the current moment robot The derivative of the output error feedback with its neighbors, For the current moment robot The derivative of the estimated value of the motion state, For the current moment robot The estimated value of the residual signal contains a mixed signal of heterogeneous information and fault information, that is, .

2. A multi-robot system fault detection method for external interference according to claim 1, characterized in that: The residual generation algorithm is to isolate the fault signal by setting the orthogonal complement space. The corresponding calculation expression is: Where, For The residual signal obtained by isolating the fault signal, For robots The orthogonal complement space of .

3. The method for detecting faults in a multi-robot system under external interference according to claim 1, wherein: The residual generation algorithm is Multiply both sides by , we get the estimated value of the balanced intermediate observer, and the corresponding calculation expression is: Where, For balance The residual signal obtained after Adjust the vector to be designed.

4. A multi-robot system fault detection method for external interference according to claim 3, characterized in that: The residual generation algorithm also sets performance indicators during processing to assist in completing the vector The design achieves a balance between fault sensitivity and robustness to heterogeneous information. The calculation expression of the performance index is: Where, For performance indicators.

5. The method for detecting faults in a multi-robot system under external interference according to claim 1, wherein: The expression of the residual evaluation rule is: Where, The current moment robot generated by the residual generation algorithm The residual result of is the threshold function.

6. A multi-robot system fault detection method for external interference according to claim 5, characterized in that: The threshold function The selection is made based on minimizing the false alarm rate while maximizing the sensitivity to failures.

7. The method for detecting faults in a multi-robot system under external interference according to claim 1, wherein: The position sensor equipped on each robot is a visual sensor or a radar.

8. The method for detecting faults in a multi-robot system under external interference according to claim 1, wherein: The multi-robot system is a heterogeneous multi-robot system or a homogeneous multi-robot system.

Images