On The Mapping Of Multi-Robot In Unknown Environment

With the development of the science, the study of robot has come into spaceflight, military and other fields from industry. There are some environment is not fit for human to enter, such as planet exploring and wasteland. Now we need robots to enter into these environments to complete some special tasks. Before this, we need get the information of these environments, and it is meaningful in application.At present, some of the work using a single robot to explore environment information, the shortage of it is: the slow pace of exploration, less aware of environmental information integrity, and system fault-tolerance and robustness is poor. However the use of multi-robot to collaborative exploring map will be increase the efficiency and accuracy. On the other hand, the problem of multi-robot mapping in unknown environment in the construction of this problem, the multi-robot systems, multi-robot research collaboration and coordination between is of great significance. Map in the Construction of a multi-robot systems with single robot system compared obvious advantages: First, a number of parallel robots can explore, explore a regional rate faster than single robot; Secondly, the robot can be more than Redundant information increased, compared with single robot system, the multi-robot systems has more fault-tolerant and more robust. However, the formulation of a strategy of multi-robot to exploring environment is more difficult to choose than a single robot exploring. In this paper, we mainly discuss with the choice of the use of coordination mechanisms for multi-robot system to complete exploring unknown static environment. Specific as follows:1. Mapping in unknown environment, the basic problem we must first solved is how to choose a method, it is limited in the future to exploring trategy. The beginning of this article we introduce some of the methods of how to describe a map: grid, topological and geometric information. According to the mandate of the follow-up has been selected to explore strategies to deal with the use of grid map. Because of the grid map is very easily to create and maintain, the robot detected by each of the grid information directly with the environment in a regional counterparts. With grid maps, can be easily completed location and path planning.2. According to various application backgrounds we distinguish the multi-robot mapping this task is divided into two kinds, based on priority of tasks and based on priority of efficiency. In this paper, multi-robot is used to detecting mines in environment, the introduction of the market based method is improve the efficiency of the system tasks. Then we analyze the factors affecting the efficiency of the task.The most important in the market based method is, with the tasks have been completed, the system of each individual in the pursuit of its own interests to maximize At the same time, the whole system is to achieve the maximum benefits and to make contributions. By calculating each individual may be taken after the decision of the proceeds and dissipation to determine the implementation of its mandate, so as to achieve the maximum net revenue system. In this paper the design of a new bid function, take into account both the individual and take care of the overall state of information, the principle behind the experiment can clearly see the feasibility and robustness.3. Research in Construction of multi-robot map mandate of the map at the same time, the macro view on the robot map Construction task groups set up a mathematical model. In a macro modeling method using probability equation, the definition adopted by the state several robot portray robot individual and the environment and other interaction between the individual circumstances. Text on multi-robot systems is not entirely the existence of the communication system in a detailed discussion. From the back of the simulation results can be seen on the model is correct. Two conclusions in the analysis method is the same.4. Based on C/S structure, distributed multiple mobile robots simulation system is developed to simulate all the functions of the real robot systems. And it is designed in the simulation system, which realized the distributed and cooperative characteristics of MRS. Later, this system is used to simulate various working environments, running in this way: the display of the environment and the supervision of the global states are completed in the server, and the client simulates other entities related to the simulation. Considering that JAVA language has the special character of platform-independence and algorithm- independence, so it is used to exploit and realize. It is practically proved that the simulation system supplies a universal platform for organizing MRS reasonably. Now we can complete some typical tasks on the platform, such as foraging, mapping, fire fighting etc.Summarily, this paper accomplishes some practicality works circumfuse with mapping: we establish a multi-robot mapping macro-model, using a market based method to optimize the algorithms in the multi-robot mapping, some algorithms are verified in the simulation system developed by this paper.