CN-121979279-A - Co-positioning and formation control method based on perception limited disturbed distance sensor

CN121979279ACN 121979279 ACN121979279 ACN 121979279ACN-121979279-A

Abstract

The invention belongs to the field of multi-agent cooperative control, in particular to a cooperative positioning and formation control method based on a perception limited disturbed distance sensor, the formation process comprises a leader and a follower, wherein the follower measures the relative distance between adjacent nodes under the constraint of the perceived radius, and the relative distance comprises the real relative distance and random noise interference. The invention considers the limited situations of limited sensing radius and random noise interference of the sensor, constructs a self-adaptive relative positioning algorithm according to the measurement of the relative distance between the follower and the adjacent node to obtain the relative position estimated value of the node at each moment, designs a distributed control law with an attenuation disturbance sequence according to the relative position estimated value, breaks through the limitation of a separate frame of 'positioning before and control after' in the prior method, and realizes the high-precision relative positioning and stable formation control in a complex environment.

Inventors

LIU YUJING
LIU ZHIXIN
LI YIBEI

Assignees

中国科学院数学与系统科学研究院

Dates

Publication Date: 20260505
Application Date: 20260116

Claims (10)

1. The co-location and formation control method based on the perception limited disturbed distance sensor is characterized in that, The team includes 1 leader A leader follows a deterministic trajectory or is located at a fixed position during the formation, each follower having a determined and bounded desired relative trajectory, the leader being arranged to Time of day, follower individuals Being able to measure and neighbor individuals The relative distance between the two comprises Time follower And neighbors The true relative distance between the two and random noise interference; the co-location and formation control method based on the perception limited disturbed distance sensor comprises the following steps: s1, according to follower And nearby individuals Relative distance between the two measurements Constructing a self-adaptive relative positioning algorithm to obtain a follower And neighbors At the position of Relative position of time of day Estimate of (2) The following are provided: In the formula, Is a follower And neighbors At the position of An estimate of the relative position of the moments in time, Is a follower And neighbors At the position of The control input difference value of the moment in time, Representation of Front of (2) Dimension vector, expressed in Initial value of time to relative position Is a function of the estimated value of (2); Representation of Front of (2) Dimension vector, expressed in Initial value of time to relative position Is a function of the estimated value of (2); Wherein, the Is at Time of day auxiliary parameter By introducing follower And neighbors The condition of relative distance measurement at past time is expected and utilized based on auxiliary parameters Constructing an optimized objective function with respect to the relative distance measurement and its predicted value squared loss, minimizing the objective function using a gradient descent algorithm for Is updated according to the update of (a); S2, according to follower And neighbors Relative position estimate of the space Designing a distributed adaptive control algorithm to achieve the desired formation performance: In the formula, Is a follower At the position of A control input of the moment of time, Is a mutually independent and bounded decay sequence, Wherein the coefficient is Is a follower And neighbors The weight of the two-way valve is equal to the weight of the two-way valve, Representing follower Is a neighbor set of (a); Is a follower At the position of The desired relative position of the moment in time with respect to the leader, Is a neighbor At the position of The desired relative position of the moment in time with respect to the leader, Is a follower At the position of The moment in time relative to the desired trajectory of the leader, Indicating that the leader is at Relative to time of day Displacement of time.
2. The method for co-locating and convoying control based on a perceptually limited disturbed distance sensor as defined in claim 1, wherein the auxiliary parameter is utilized An adaptive projection operator is introduced to ensure a good definition of the algorithm.
3. The method for co-locating and convoying control based on perceptually limited disturbed distance sensor as defined in claim 2, wherein a weighted adaptive sequence is introduced to ensure the auxiliary parameters The finite nature of the estimation and the stability of the closed loop system.
4. The method of co-location and formation control based on perceptually limited disturbed distance sensors as defined in claim 3, wherein relative distance measurements are utilized Lower bound property of condition expectations to ensure auxiliary parameters Convergence of the estimate.
5. The co-location and formation control method based on a perceptively limited disturbed distance sensor as claimed in claim 4, characterized in that said formation is unmanned aerial vehicle formation or unmanned vehicle formation.
6. The method for co-location and formation control based on perceptually limited disturbed distance sensors as defined in claim 3, wherein said patterns of formation include relative parking, circular motion, and surrounding of static or dynamic objects.
7. The perception limited disturbed distance sensor-based co-location and formation control method as claimed in claim 6, wherein the follower Capable of receiving neighbors Is a piece of information of (a).
8. The perception limited disturbed distance sensor-based co-location and formation control method of claim 7 wherein the leader and follower each follow a first order discrete time dynamic model: Wherein the method comprises the steps of Is a follower At the position of The position of the moment in time, Is a follower At the position of A control input of the moment of time.
9. The method for co-locating and convoying control based on the perceptually limited disturbed distance sensor as defined in claim 8, wherein, in Time of day, follower Measured and neighbor The relative distance between them is: Wherein the method comprises the steps of Follower obtained for measurement And neighbors The relative distance between the two plates is set to be equal, Representation of Time follower And neighbors The true relative distance between the two, Representative of The random noise interference at the moment in time, Is a saturation function.
10. The method for co-locating and convoying control based on a perceptually limited disturbed distance sensor as defined in claim 9, The definition is as follows: Wherein the method comprises the steps of Is a follower Is included in the model (a).

Description

Co-positioning and formation control method based on perception limited disturbed distance sensor Technical Field The invention belongs to the field of multi-agent cooperative control, and provides a method for realizing distributed cooperative positioning and formation control based on relative distance measurement, which is particularly suitable for complex task scenes with failure of global positioning signals, limited local perception radius and random noise interference measurement. Background In the field of autonomous system and multi-agent cooperation, distributed formation control is used as a core technology for realizing efficient cooperation of clusters, and is widely applied to numerous actual scenes such as unmanned vehicle clusters, unmanned aerial vehicle formation, mobile robots and the like through a basic paradigm of 'local perception-information interaction-cooperative control'. Most of the existing formation control methods rely on external infrastructures such as a Global Positioning System (GPS) to provide accurate absolute/relative position information of an intelligent agent, and have serious defects in adaptability and reliability in GPS rejection environments such as severe weather, electromagnetic interference, hardware faults and the like and complex dynamic environments with random interference. In order to realize efficient coordination of multi-agent systems in complex dynamic environments, development of a co-location and formation control technology based on distance and other airborne sensors is needed. The method is characterized in that the relative positions among the intelligent agents are estimated in real time by utilizing the distance measurement information among the intelligent agents, and distributed closed-loop control is designed based on the relative positions, so that a preset formation target is finally achieved. Aiming at the relative positioning problem based on distance measurement, the prior art generally adopts an inverse function method to linearize a nonlinear distance measurement model, combines classical estimation algorithms such as least square, least mean square and the like to design a distributed positioning algorithm, realizes the cooperative positioning among intelligent agents and is applied to formation control. However, in a practical complex scenario, the distance sensor often has inherent constraints of limited sensing radius, easy random noise interference of measurement and the like, and at this time, the problem of co-location among multiple agents is essentially a non-convex optimization problem, which makes it difficult for the traditional positioning algorithm to ensure convergence of estimation errors and more difficult to provide formation control performance guarantee. Therefore, under the constraint of the coexistence of the perception limitation and the random noise interference, how to design a relative positioning algorithm based on distance measurement to realize high-precision relative positioning among multiple intelligent agents, and further design a controller on the basis to realize accurate formation control of the multiple intelligent agent system, thereby having important engineering application value and urgent practical application requirements. However, due to the strong nonlinear coupling between the distance measurement and the relative position, the positioning target and the control task are mutually restricted, and the design of the distributed co-positioning and control integrated algorithm faces serious challenges. How to break through the core bottleneck that the existing algorithm has strict requirements on the sensing performance of the sensor and the adaptability of the complex dynamic environment is insufficient, and the construction of the integrated framework for the co-location and formation control with strong adaptability and high precision is a key technical direction of urgent breakthrough in the field. In the process of implementing the present invention, the inventor finds that at least the following problems exist in the prior art: (1) Most of the existing methods adopt a separate framework of 'positioning first and control later', require that the system track meets strict conditions such as continuous excitation, neglect the dynamic coupling relation between positioning estimation and closed-loop control, and are difficult to be suitable for complex dynamic self-adaptive scenes. (2) The existing distributed co-location algorithm based on distance measurement is mostly built under a deterministic system framework, and random noise interference commonly existing in sensor measurement is not considered, so that algorithm performance in an actual random interference environment is lack of convergence guarantee. (3) When the existing research designs a positioning and formation control algorithm, the actual physical constraints of limited sensing radius, non-negative measurement value and the l