Research On Essential Techniques For Mobile Intelligent Robot And Robot Team

With developing on the intelligent robot technology, people have been noticed the importance of cooperating among robots, and the challenge in practice. After researching foundations on the single robot had been established, the studies for theories and techniques on robot team have been developed. In this thesis, the multi-agent system, especially for several key techniques of autonomous intelligent mobile robot are discussed in detail. Some related models and schemes of realization are presented.It is first introduced about the researches on autonomous intelligent mobile robot (A1MR), and emphasized on multi-agent system (MAS). A summary is made on the developments of AIMR both in the form of individual and a group based on MAS. And main research directions in the recent future are proposed.The architecture of mobile robot is one of the key points. A proper architecture is fundamental for the intelligent robot system construction according to the different tasks. Considering the characteristics of the mobile robots, some researches have been done with the principle of layered function. Relying on the MAS describing capability, a three layers structure model, which is suitable for mobile robot systems, is also presented here. That is, organization layer, relation layer, individual layer. Each layer's detail combined with an example is given.Environment perception is a key technique for mobile robots. It primarily includes two-dimension and three-dimension information processing and understanding. The former deals with road recognition; the latter is the main means of terrain understanding and obstacle detecting. A noise-suppressing approach for range data, based on Markov Random Field (MRF), is proposed, which provides an effective method for range data pre-processing.Data fusion and local path planning is the embodiment of mobile robot intelligent behaviors on the middle and high level. The former is used to understand the environment correctly by synthesizing all information gathered and processed from several kinds of sensors; the latter is used to describe an action for robots based on global aim and understanding environment. Therefore, efforts on the following problems are made. First, some problems for the calibration of line-array laser radar are solved, such as template design and feature point acquiring. To solve the problems caused by radar installation and lack of information, which inevitably leads to false report at a relatively high rate and a radar blind area in front of the vehicle, a data-fusion model is designed for obstacle detection and memory. Finally, for the local path planning, an anticipant path following, rolling and adjusting plan are used here. A satisfied performance is obtained by solving the problem balanced between the decision time and the error disturbing. Therefore, the smoothing path can be guaranteed.A hierarchic method for speed control of mobile robots is proposed and designed, which means that five control levels are defined, these are target speed control, state speed control, action speed control, performing speed control and low level speed control. By relevant speed control knowledge, the whole performance of speed control is improved.Research on robot team is a developing and novel discipline. The following works are finished in this field. First, a unified description based on agents for robots is given; the concepts of agent, from the microcosmic, the macrocosm!c, to the society, are defined. Then an algorithm for robot society construction is designed.By analyzing the cooperating form among robots and current technical development, a communication strategy of the robot is discussed. The point-to-point model is adopted within a team. And a team leader, acting as a representative, communicates with other team leaders also in the same way. That means, a robot in a team can't communicate directly with other robots in a different team.A new method is brought forward on multi agent alliance. The conceptIVof the skill ability matri...