scholarly journals A Review of Mobile Robot Path Planning Based on Deep Reinforcement Learning Algorithm

2021 ◽  
Vol 2138 (1) ◽  
pp. 012011
Author(s):  
Yanwei Zhao ◽  
Yinong Zhang ◽  
Shuying Wang
Keyword(s):  
Deep Learning ◽  
Reinforcement Learning ◽  
Path Planning ◽  
Mobile Robot ◽  
Video Game ◽  
Autonomous Navigation ◽  
Learning Algorithm ◽  
Basic Knowledge ◽  
Target Point ◽  
Reinforcement Learning Algorithm

Abstract Path planning refers to that the mobile robot can obtain the surrounding environment information and its own state information through the sensor carried by itself, which can avoid obstacles and move towards the target point. Deep reinforcement learning consists of two parts: reinforcement learning and deep learning, mainly used to deal with perception and decision-making problems, has become an important research branch in the field of artificial intelligence. This paper first introduces the basic knowledge of deep learning and reinforcement learning. Then, the research status of deep reinforcement learning algorithm based on value function and strategy gradient in path planning is described, and the application research of deep reinforcement learning in computer game, video game and autonomous navigation is described. Finally, I made a brief summary and outlook on the algorithms and applications of deep reinforcement learning.

scholarly journals Autonomous Path Planning Scheme Research for Mobile Robot

2016 ◽  
Vol 16 (4) ◽  
pp. 113-125
Author(s):  
Jianxian Cai ◽  
Xiaogang Ruan ◽  
Pengxuan Li
Keyword(s):  
Path Planning ◽  
Mobile Robot ◽  
Autonomous Navigation ◽  
Learning Algorithm ◽  
Action Learning ◽  
Cognitive Learning ◽  
Learning Ability ◽  
Q Learning ◽  
Planning Strategy ◽  
Navigation Path

Abstract An autonomous path-planning strategy based on Skinner operant conditioning principle and reinforcement learning principle is developed in this paper. The core strategies are the use of tendency cell and cognitive learning cell, which simulate bionic orientation and asymptotic learning ability. Cognitive learning cell is designed on the base of Boltzmann machine and improved Q-Learning algorithm, which executes operant action learning function to approximate the operative part of robot system. The tendency cell adjusts network weights by the use of information entropy to evaluate the function of operate action. The results of the simulation experiment in mobile robot showed that the designed autonomous path-planning strategy lets the robot realize autonomous navigation path planning. The robot learns to select autonomously according to the bionic orientate action and have fast convergence rate and higher adaptability.

Genetic Q-Fuzzy Based Intelligent Control for Mobile Robot Navigation

2004 ◽  
Author(s):  
V. Ram Mohan Parimi ◽  
Devendra P. Garg
Keyword(s):  
Genetic Algorithm ◽  
Fuzzy Logic ◽  
Reinforcement Learning ◽  
Mobile Robot ◽  
Fuzzy Logic Controller ◽  
Learning Algorithm ◽  
Robot Navigation ◽  
Path Tracking ◽  
Mobile Robot Navigation ◽  
Reinforcement Learning Algorithm

This paper deals with the design and optimization of a Fuzzy Logic Controller that is used in the obstacle avoidance and path tracking problems of mobile robot navigation. The Fuzzy Logic controller is tuned using reinforcement learning controlled Genetic Algorithm. The operator probabilities of the Genetic Algorithm are adapted using reinforcement learning technique. The reinforcement learning algorithm used in this paper is Q-learning, a recently developed reinforcement learning algorithm. The performance of the Fuzzy-Logic Controller tuned with reinforcement controlled Genetic Algorithm is then compared with the one tuned with uncontrolled Genetic Algorithm. The theory is applied to a two-wheeled mobile robot’s path tracking problem. It is shown that the performance of the Fuzzy-Logic controller tuned by Genetic Algorithm controlled via reinforcement learning is better than the performance of the Fuzzy-Logic controller tuned via uncontrolled Genetic Algorithm.

Deep Deterministic Policy Gradient for Navigation of Mobile Robots

2021 ◽  
Vol 40 (1) ◽  
pp. 349-361
Author(s):  
Junior Costa de Jesus ◽  
Jair Augusto Bottega ◽  
Marco Antonio de Souza Leite Cuadros ◽  
Daniel Fernando Tello Gamarra
Keyword(s):  
Reinforcement Learning ◽  
Mobile Robot ◽  
Learning Algorithm ◽  
Mobile Robot Navigation ◽  
Simulated Environments ◽  
The Neural Network ◽  
Policy Gradient ◽  
Reward Functions ◽  
Continuous Actions ◽  
Reinforcement Learning Algorithm

This article describes the use of the Deep Deterministic Policy Gradient network, a deep reinforcement learning algorithm, for mobile robot navigation. The neural network structure has as inputs laser range findings, angular and linear velocities of the robot, and position and orientation of the mobile robot with respect to a goal position. The outputs of the network will be the angular and linear velocities used as control signals for the robot. The experiments demonstrated that deep reinforcement learning’s techniques that uses continuous actions, are efficient for decision-making in a mobile robot. Nevertheless, the design of the reward functions constitutes an important issue in the performance of deep reinforcement learning algorithms. In order to show the performance of the Deep Reinforcement Learning algorithm, we have applied successfully the proposed architecture in simulated environments and in experiments with a real robot.

DEEP-SARSA: A REINFORCEMENT LEARNING ALGORITHM FOR AUTONOMOUS NAVIGATION

2001 ◽  
Vol 12 (10) ◽  
pp. 1513-1523 ◽  
Author(s):  
M. ANDRECUT ◽  
M. K. ALI
Keyword(s):  
Reinforcement Learning ◽  
Autonomous Navigation ◽  
Learning Algorithm ◽  
Robot Navigation ◽  
Learning Algorithms ◽  
Delayed Reinforcement ◽  
Graph Searching ◽  
Ill Posed ◽  
Autonomous Robot Navigation ◽  
Reinforcement Learning Algorithm

In this paper we discuss the application of reinforcement learning algorithms to the problem of autonomous robot navigation. We show that the autonomous navigation using the standard delayed reinforcement learning algorithms is an ill posed problem and we present a more efficient algorithm for which the convergence speed is greatly improved. The proposed algorithm (Deep-Sarsa) is based on a combination between the Depth-First Search (a graph searching algorithm) and Sarsa (a delayed reinforcement learning algorithm).

Sign in / Sign up

Export Citation Format

Share Document