| This paper presents a mobile robot system with sub-tracks for search and rescue. There are often natural disasters and accidents occurred in China, such as coal mine accidents, mud flow, hurricane and so on. It brings great danger to the rescue members because of the unknown and complex environment, and the potential secondary damage, the second coal mine explosion, the second building collapse, for example. So it is necessary to develop a substitute for the rescue members to detect the dangerous areas after the disasters, which can provide useful information of the environment, and strong support for the planning and scheduling for the rescue operation.This paper deals with the basic control technology at the beginning, which is the basis of the stable and effective operation. With the improvement of our search and rescue robot platform from LUKER-I to LUKER-II, we construct the distributed control network instead of the centralized control structure. The implementation of distributed network based on CAN bus is given in detail. The basic CAN notes which play a key role in the system, the motion control node, the pose node, and the multifunction node, for example, are designed in detail in this paper. This part lay the foundation for the following research of the intelligent behaviors.A human-robot collaborative framework is established in the tele-operation mode. In this framework, the operator and robot take full advantage of their own strengths while collaborating quite closely with each other, in order to improve the efficiency of the overall implementation of the task. On the one hand, the operator can control the robot's movement directly known as the direct control mode. On the other hand, the operator can simply give a task to the robot through the interactive interface, and the robot can take task decomposition and implement behavior control intelligently. In the process the operator can monitor the real-time task implementation. If something unexpected occurred, the robot can be controlled by the operator in time. The behavior-based control system in the collaborative framework is studied in this paper as well. We design a reflective behavior in order to provide pose protection of the robot in the unstructured environment, which make up one of the behaviors in parallel. The pose reset task is intelligently controlled with the method of FSM, in which the pose protection is involved in the framework of human-robot collaborative mode.In this paper, a detailed task namely the obstacle climbing is given in the collaborative framework. Two different motion sequences are designed to climb a square obstacle. The maxim height of the obstacle and the appropriate angle of sub-tracks can be derived from the analysis of the robot Centroid. The task is programmed into two stages in the plan layer, namely the pre-posture and the climbing. The pre-posture adjusts the robot pose with the method of FSM. The climbing adopts the fuzzy Q learning to maintain the stability of the process. In order to improve the real time response, the control law derived from the fuzzy Q learning are put into the behavior layer to execute. In addition the pose protection behavior is executed simultaneously. The process of obstacle climbing can be monitored in the supervision layer as well.This paper presents the human-robot interactive interface for the search and rescue robot. The 3D models of the robot is built with the 3D studio software and can be rebuilt in the Visual C++ with OpenGL programming. It improves the human-robot interface friendly.The paper studies some key technologies for the development of the search and rescue robot with the trend of intelligent control. The experiments and tests in the underground coal mine proved the invalidation of the research. The research can help improve the practical use of the specialized robot for search and rescue, such as the fire detect robot, or other special operating environment disaster rescue robot used for earthquake, landslides, and hurricanes. |
PDF Full Text Request