Multi-robot collaborative navigation and obstacle avoidance method

A collaborative navigation and multi-robot technology, applied in the field of robot navigation, can solve problems such as the amount of calculation can no longer support high real-time requirements, system stability deterioration, etc., to reduce training time, solve difficult convergence, and speed up the training process Effect

Active Publication Date: 2021-12-21
SUN YAT SEN UNIV
View PDF9 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] An important prerequisite for the realization of these intelligent applications is that the robot has a strong ability to avoid obstacles in an unknown dynamic environment. However, in this environment, the mobile robot cannot obtain the position information of obstacles or other robots, and can only observe the local information acquired by its own sensors. At this time, the huge amount of calculation generated by the traditional obstacle avoidance algorithm of "building a map first and then planning a path" can no longer support the application to meet the high real-time requirements of the 5G era, and changes in the number and position of obstacles will also cause system poor stability

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-robot collaborative navigation and obstacle avoidance method
  • Multi-robot collaborative navigation and obstacle avoidance method
  • Multi-robot collaborative navigation and obstacle avoidance method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0073] This embodiment includes a network of actors and a network of judges, such as Figure 5 and Figure 6 . The actor network structure mainly includes convolutional layers, maximum pooling layers, linear layers, activation layers, and long-short-term memory layers. The activation function uses a linear correction unit activation function to solve the problem of gradient disappearance and neuron "death". The convolutional layer and the pooling layer process the input red, green and blue three-channel image to extract features. The judge network adds action variables in the input of long and short-term memory, and the others are consistent with the actor network.

[0074] The convolutional neural network consists of two convolutional layers, the convolution kernel sizes are 16×3×3 and 32×3×3, the number of channels is 16 and 32, and the step size is 3. The maximum pooling layer selects the maximum value in the convolution kernel as output.

[0075] The hyperparameters ar...

Embodiment 2

[0078] The purpose of this embodiment is to implement a deep reinforcement learning algorithm network for a single robot to learn obstacle and reward information in the environment. Figure 8 It is an implementation flowchart of a deep reinforcement learning algorithm of a long short-term memory-deep deterministic policy gradient algorithm for multi-robot navigation provided by an embodiment of the present invention. As shown in the figure, the method may include the following steps:

[0079] S1: Establish the state model of the robot: use partially observable Markov decision process to infer the distribution of the state of the robot according to the observation information of the environment, and describe it with a six-tuple (S, A, T, R, Z, O).

[0080] S2: Establish a convolutional neural network for robot camera data processing: obtain the current camera data of the robot, perform Gaussian blur and scale transformation, and obtain the vector of the robot's image observation...

Embodiment 3

[0127] Embodiment 3 of the present invention provides an improved long-short-term memory network. This embodiment adopts the network structure of long-short-term memory-deep deterministic policy gradient algorithm. The purpose is to introduce short-term memory, speed up and improve the speed and performance of the reinforcement learning training process , adapted to a reinforcement learning process with long short-term memory networks. The improvement part mainly includes two parts: random update and jump update. image 3 It is a schematic diagram of the random update of the long short-term memory network of a specific embodiment of the multi-robot cooperative navigation and obstacle avoidance method and system in the present invention. As shown in the figure, the strategy randomly selects a time point in each randomly selected overall strategy , start learning with a fixed number of steps from this moment, and reset the state of the long-short-term memory hidden layer to zero...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a multi-robot collaborative navigation and obstacle avoidance method. The method comprises the following steps of modeling a decision process of a robot in an unknown environment according to a partially observable Markov decision process; according to the environment modeling information of the current robot, introducing a depth deterministic strategy gradient algorithm, extracting a sampled image sample, and inputting the sampled image sample into a convolutional neural network for feature extraction; improvement being carried out on the basis of a depth deterministic strategy gradient algorithm, a long-short-term memory neural network being introduced to enable the network to have memorability, and image data being more accurate and stable by using a frame skipping mechanism; and meanwhile, an experience pool playback mechanism being modified, and a priority being set for each stored experience sample so that few and important experiences can be more applied to learning, and learning efficiency is improved; and finally, a multi-robot navigation obstacle avoidance simulation system being established. The method is advantaged in that the robot is enabled to learn navigation and obstacle avoidance from easy to difficult by adopting a curriculum type learning mode so that the training speed is accelerated.

Description

technical field [0001] The invention relates to the field of robot navigation, in particular to a multi-robot cooperative navigation and obstacle avoidance method. Background technique [0002] With the maturity of 5G technology, robot technology has entered human life and work in all aspects, such as automatic driving, automatic transportation, search and rescue, etc. Due to the sharp increase in the demand for human manufacturing applications, especially for the small-batch and multi-variety personalized production requirements in intelligent manufacturing, in response to this complex and flexible production trend, the operation function of a single robot begins to appear relatively simple, and production needs to be more digitized. Networking and intelligence make it inevitable that the theory and application of multi-robots will develop. Multi-robot collaboration can complete processing more accurately and efficiently and reduce consumption. For example, in processing a...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G05D1/02
CPCG05D1/0221G05D1/0246G05D1/0287G05D2201/0217
Inventor 彭键清陈诺陈畅
Owner SUN YAT SEN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap