Unlock instant, AI-driven research and patent intelligence for your innovation.

Robot degenerated environment obstacle avoidance method based on impulsive neural network internal plasticity

A spiking neural network and plasticity technology, applied in the field of brain-like robots, to achieve the effect of improving the success rate

Active Publication Date: 2022-02-11
DALIAN UNIV OF TECH
View PDF6 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, no work has yet applied dynamic spiking thresholds with a biological theoretical background to practical tasks based on spiking neural networks, so designing biologically inspired homeostatic models of intrinsic plasticity is an urgent task in the field of biological brain-inspired robotics

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
  • Robot degenerated environment obstacle avoidance method based on impulsive neural network internal plasticity
  • Robot degenerated environment obstacle avoidance method based on impulsive neural network internal plasticity
  • Robot degenerated environment obstacle avoidance method based on impulsive neural network internal plasticity

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The specific implementation manners of the present invention will be further described below in conjunction with the drawings and technical solutions.

[0039] This method uses LIF and SRM neuron models as the main neuron structure of the network, and uses DDPG as the framework of deep reinforcement learning. The state includes lidar data, the distance from the target point and the speed at the last moment; the action is composed of brain-like robots. Composed of linear velocity and angular velocity; the reward function includes the state of the distance from the target at each moment (the closer is the positive reward, and vice versa), it is -20 if there is a collision, and 30 if it reaches the target point, encouraging the robot to take every step The range of action taken should not be too large, that is, it should not exceed 1.7 times the angular velocity at the previous moment.

[0040] The reinforcement learning algorithm is implemented in Pytorch. Stochastic gra...

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 robot degraded environment obstacle avoidance method based on impulsive neural network inherent plasticity. The method comprises a dynamic energy threshold module, a dynamic time threshold module, a biological reasonable dynamic energy-time threshold fusion module and a synaptic scene building and autonomous learning module. A decision-making network in the synaptic autonomous learning module takes laser radar data, the distance to a target point and the speed at the last moment as state input of the decision-making network, through autonomous adjustment of a dynamic energy-time threshold value, the speed of a left wheel and the speed of a right wheel of the robot are output, and therefore autonomous sensing and decision making are carried out. The invention solves the problem that a pulse neural network lacks internal plasticity, so that a model is in steady-state imbalance and is difficult to adapt to a degenerated environment, and is successfully deployed in a mobile robot to maintain a stable triggering rate, so that autonomous navigation and obstacle avoidance in degenerated, interference and noise environments are performed, and the method has effectiveness and applicability in different degradation scenes.

Description

technical field [0001] The invention belongs to the field of brain-like robots in the field of brain-like intelligence, and the specific realization result is autonomous navigation and obstacle avoidance of brain-like robots, and particularly relates to an obstacle avoidance in a degraded environment with intrinsic plasticity and stability of impulse neurons method. Background technique [0002] The robot obstacle avoidance task is that in a more complex scene, the robot can autonomously navigate to the target point without any collision with the obstacle, which has great practical application value. With the rapid development of artificial intelligence technology, tasks related to robot obstacle avoidance, such as sweeping robots, unmanned driving, smart warehouses, smart logistics, etc., have achieved significant performance improvements. [0003] Although some artificial neural network-based methods have been successfully applied to obstacle avoidance tasks, their high e...

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): G06N3/04G06N3/08G06F17/10
CPCG06N3/049G06N3/08G06F17/10G05D1/0221G05D1/024G06N3/092B25J9/161B25J9/163B25J9/1666B25J9/1671
Inventor 丁建川杨鑫董博尹宝才周运铎王洋
Owner DALIAN UNIV OF TECH