CN-121985306-A - Indoor robot positioning method based on multi-description coding-decoding scheme

CN121985306ACN 121985306 ACN121985306 ACN 121985306ACN-121985306-A

Abstract

The application provides an indoor robot positioning method based on multi-description coding-decoding, which is technically characterized by comprising the following steps of establishing a kinematic model of an indoor robot and a measurement model of a sensor, and establishing a mathematical model; and solving the upper bound of the positioning error covariance, optimally designing the upper bound of the positioning error covariance minimized by the parameters of the estimator, and realizing the accurate positioning of the indoor robot. The application actively adjusts the transmission time of the sensor data through the token bucket communication protocol, effectively suppresses the burst flow, avoids the instantaneous overload of the channel, realizes the high-efficiency and smooth utilization of communication resources, adopts a multi-description coding-decoding scheme to divide each measurement information into two independent descriptions and transmits the two independent communication links in parallel, and even if partial descriptions are lost due to insufficient tokens or bad channels, the decoder can still effectively reconstruct based on the single description, thereby improving the fault tolerance of the indoor robot positioning system.

Inventors

GAO RUIFENG
Qi Qingchi
HUANG CONG
SHI QUAN

Assignees

南通大学

Dates

Publication Date: 20260505
Application Date: 20260112

Claims (10)

1. The indoor robot positioning method based on the multi-description coding-decoding scheme under the token bucket communication protocol is characterized by comprising the following steps: s1, establishing a kinematic model of the indoor robot and a measurement model of a sensor, wherein the kinematic model is used for showing the relation between the state of the indoor robot and control input, and the measurement model is used for showing the measurement relation between the sensor and landmark distance and phase angle of the indoor robot; S2, establishing a mathematical model for representing a token bucket communication protocol and a multi-description coding-decoding scheme, wherein the token bucket communication protocol controls the transmission time of sensor data by dynamically updating the number of tokens in a bucket, and only allows data transmission when the number of tokens in the bucket meets a transmission threshold; S3, constructing a state estimator of the indoor robot based on the measurement signals reconstructed by the decoder; and S4, solving a positioning error covariance upper bound, and minimizing the positioning error covariance upper bound through optimizing design of an estimator parameter to realize accurate positioning of the indoor robot.
2. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, wherein the method for setting a kinematic model of the robot in S1 is as follows: Is provided with Is a state vector of the indoor robot, For control input of the indoor robot, the indoor robot system may write: Wherein the method comprises the steps of Indicating the position of the indoor robot, Is azimuth; And (3) with In order to achieve both the displacement speed and the angular speed, For the sampling period of the indoor robot, Zero mean Gaussian white noise, covariance 。
3. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, wherein the measurement model of the sensor in S1 is as follows: in the above-mentioned formula(s), Is a Gaussian white noise, and the covariance is expressed as ; Is a measurement of the sensor in an ideal state, defined as follows: Wherein the method comprises the steps of Is an indoor robot to landmark Is used for the distance of (a), Is the phase angle.
4. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, wherein the token bucket communication protocol in S2 is as follows The update rule of (2) is given by the following formula: in the above-mentioned formula(s), Representation of Number of tokens in time bucket, and Satisfy the following requirements , Is shown in The number of tokens added to the bucket at the moment Is the maximum capacity of the bucket; Is shown in The number of tokens that need to be consumed to transmit data at the moment, A variable is indicated for the transmission license.
5. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, wherein the S2 uses the multiple description coding-decoding scheme to measure signals received by an encoder Processing is performed for each measurement component Two description signals are generated by two independent coding functions respectively: 。
6. the method for indoor robot positioning based on multiple description coding-decoding scheme according to claim 5, wherein the decoding rule of the decoder is as follows: in the above-mentioned formula(s), And Is an auxiliary decoder, uses only a single description for reconstruction, Is a central decoder that achieves higher accuracy reconstruction by combining the two descriptions and that maintains the output of the previous moment if both descriptions are lost.
7. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, wherein the estimator constructed in S3 is as follows: Wherein the method comprises the steps of As the parameters of the estimator to be designed, Is an innovative function, defined as follows: 。
8. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, wherein the upper bound matrix of positioning error covariance in S4 And The following recurrence relation is satisfied: And Wherein the initial conditions are Positive scalar quantity And The method meets the following conditions: Wherein the method comprises the steps of Then Is that Is an upper bound of (a) 。
9. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, wherein the local estimator parameter in S4 is The requirements are as follows: Wherein the method comprises the steps of In the form of a covariance correlation matrix, For observing the correlation matrix.
10. The method for positioning an indoor robot based on a multiple description coding-decoding scheme according to claim 1, further comprising the step of 5 verifying the validity of the provided positioning algorithm and performing an experiment-based verification: the verification scheme comprises the following steps: step 5-1, constructing a simulation experiment platform; Step 5-2, setting basic parameters; Step 5-3, 1) calculating estimator parameters according to the formula in S4 2) Calculating the positioning information in the next step according to the formula in S3 Then calculating the upper bound of the positioning error covariance according to the formula in S4 And back to 1) until the end; step 5-4, adopting a recursive least square method, wherein the final evaluation standard of the positioning effect is shown in the following formula Wherein Is the mean square error.

Description

Indoor robot positioning method based on multi-description coding-decoding scheme Technical Field The application relates to the technical field, in particular to an indoor robot positioning method based on a multi-description coding-decoding scheme. Background In recent years, the rapid development of technologies such as edge computing, 5G communication and artificial intelligence provides a powerful technical support for the wide application of indoor robots. If the indoor robot can provide effective service in the scenes of logistics, inspection, home and the like, the most critical point is that the indoor robot can have high-precision and high-robustness positioning capability. However, the indoor robot is actually often in an environment complex and a communication bandwidth limited operation scene. If the sensor transmits data too frequently, the communication network is easy to be jammed, and thus the phenomenon of data loss or delay is caused. This can seriously affect the continuity and reliability of the positioning of the indoor robot. To save limited communication resources, the industry often uses various communication resources to adjust the sensor's data transmission timing. The dynamic event triggering mechanism determines whether the measurement data can be transmitted or not by setting a triggering threshold value, so that the purpose of saving communication resources is achieved to a certain extent. However, the transmission time under the dynamic event triggering mechanism has strong randomness and passivity, and burst dense transmission is easy to generate under the condition of high uncertainty of an external environment, so that an instantaneous channel overload phenomenon is caused. Especially when parallel transmission strategies such as multi-description coding are employed, multiple descriptions may be sent centrally due to synchronous triggering, resulting in transient channel overload phenomena. Once the transient channel overload phenomenon occurs, the total description is lost, thereby affecting the working performance of the indoor robot. Disclosure of Invention The application aims to solve the technical problem that the transmission moment of a dynamic event trigger mechanism in the prior art is easy to generate sudden dense transmission under the condition of higher uncertainty of an external environment, thereby causing the phenomenon of instantaneous channel overload. The indoor robot positioning method based on the multi-description coding-decoding scheme under the token bucket communication protocol is characterized by comprising the following steps: s1, establishing a kinematic model of the indoor robot and a measurement model of a sensor, wherein the kinematic model is used for showing the relation between the state of the indoor robot and control input, and the measurement model is used for showing the measurement relation between the sensor and landmark distance and phase angle of the indoor robot; S2, establishing a mathematical model for representing a token bucket communication protocol and a multi-description coding-decoding scheme, wherein the token bucket communication protocol controls the transmission time of sensor data by dynamically updating the number of tokens in a bucket, and only allows data transmission when the number of tokens in the bucket meets a transmission threshold; S3, constructing a state estimator of the indoor robot based on the measurement signals reconstructed by the decoder; and S4, solving a positioning error covariance upper bound, and minimizing the positioning error covariance upper bound through optimizing design of an estimator parameter to realize accurate positioning of the indoor robot. Preferably, the method for setting the kinematic model of the robot in S1 is as follows: Is provided with Is a state vector of the indoor robot,For control input of the indoor robot, the indoor robot system may write: Wherein the method comprises the steps of Indicating the position of the indoor robot,Is azimuth; And (3) with In order to achieve both the displacement speed and the angular speed,For the sampling period of the indoor robot,Zero mean Gaussian white noise, covariance。 Preferably, the measurement model of the sensor in S1 is as follows: in the above-mentioned formula(s), Is a Gaussian white noise, and the covariance is expressed as;Is a measurement of the sensor in an ideal state, defined as follows: Wherein the method comprises the steps of Is an indoor robot to landmarkIs used for the distance of (a),Is the phase angle. Preferably, the token communication protocol in S2 is as followsThe update rule of (2) is given by the following formula: in the above-mentioned formula(s), Representation ofNumber of tokens in time bucket, andSatisfy the following requirements,Is shown inThe number of tokens added to the bucket at the momentIs the maximum capacity of the bucket, b is expressed inThe number of tokens that need to be consumed to