Environment Exploration And Map Building For Mobile Robot Based On Biological Immune Theory

To perform the complex tasks better and more efficient, immune system is applied in robotic fields for its identifiability, dynamics and adaptivity. According to large-scale unknown environment, this thesis proposes an environmental exploration strategy for mobile robot based on immune theory. Firstly, possible nodes are abstracted using the generated admissible space tree for path planning. Then, the abstracted geometric features are related to the corresponding topology nodes to improve the environment recognition of robot, and the hybrid map is constructed based on hierarchical mixed information. On this basis, immune system is introduced. Compared with immune system, the abstracted possible nodes choosing process is optimized to improve exploration efficiency. Finally, multi-robotics system is introduced to integrate with immune system, so that multi-robotics system collaboration exploration is accomplished. The main research works are as follows:1. Hybrid map model construction.The hybrid map model based on hierarchy mixed features is constructed. Different maps are applied for different tasks: Geometry map is mainly used to improve environment recognition of robot, topology map is mainly applied for location, possible nodes map is utilized for exploration, etc. During robot environment recognition process, it needs many kinds of map information. Therefore, with hybrid map construction, the robot can seek for appropriate map information to improve the robot efficiency.2. Feature abstraction and map construction.The admissible space tree is abstracted to get possible nodes for path planning and the map is constructed in real time; the environment is divided into regular area and irregular area, Line segment algorithm and Grid-based shared Nearest Neighbor method(GNN) are introduced to abstract geometric detailed features respectively. Then, bayesian formula is adapted to process dynamic information to eliminate the influence on map construction caused by dynamic information. So the map accuracy is improved.3. Single robot exploration system based on immune theory construction.The immune system is introduced in exploration system for mobile robot. Compared with immune system, the robot is represented as B-cell, local environment information obtained from laser sensor is taken as antigen and the generated possible nodes are represented as antibody. The antibody-antigen affinity is calculated to evaluate each possible node. As a result, the optimization of exploration path is guaranteed and the exploration efficiency is improved.4. Multi-robotic exploration system based on immune theory construction.To make the robot serve the task better with higher efficiency, the multi-robotic system and immune system are combined. Compared with regulatory mechanism in immune system, interaction between antibodies is utilized to control robot action. The antibodies are chosen by calculating the stimulating parameters. In this way, the repeated exploration areas between robots will be reduced and the efficiency will be improved.The validity and practicability of the proposed environment exploration and map building approach are validated by a lot of experiments in the real experimental environment. The constructed mobile robot exploration system and map building approach can be used for real unknown environment exploration task and have a certain theoretical and practical value.