The simple act of moving, quite easy for humans, but in terms of the robot becomes extremely complicated, would have to mention when it comes to mobile robot path planning, path planning of mobile robot navigation is the most basic aspects, referring to the there are obstacles in the robot work environment, how to find a path from start to finish appropriate motion path, the robot during exercise can be safe, collision-free bypass all obstacles. This differs from the shortest path with dynamic programming and other methods to obtain, but to move the robot can make a comprehensive judgment on the static and dynamic environment, intelligent decisions.
Overall, path planning mainly involves three big problems: ① clear starting point and end point position; ② avoid obstacles; ③ achieve optimization on the path as much as possible.
Robot path planning of global and local points of planning
Depending on the degree of mastery of environmental information, path planning can be divided into global path planning and local path planning.
Global path planning in a known environment, the robot path planning, path planning accuracy depends on the accuracy of the acquired environment, global path planning can find the optimal solution, but need to know in advance the exact information environment, when the environment when changes, such as an unknown obstacles, which can not do anything. It is a kind of advance planning, so the robot system is less demanding real-time computing capabilities, although the planning result is global, and more, but the environment noise model errors and poor robustness.
The local path planning of the environmental information or some completely unknown known, focuses on the consideration of the current local environmental information robot, so the robot has a good obstacle avoidance ability to detect the working environment of the robot via sensors to obtain the location of obstacles and geometrical properties and other information, such planning need to gather environmental data, and this can be corrected at any time dynamically updated environment model, local modeling and programming method will search integration environment, the robot system requires high-speed processing capability information and computing power, there are errors on the environment and higher noise robustness, capable of real-time feedback and correction of planning results, but due to lack of information on the global environment, the planning results may not be optimal, and may even find the right or full path.
Local and global path planning and path planning is no essential difference, many global path planning method is applicable to the improved local path planning may also be used, and a method suitable for local path planning after the same is also applicable to improving global path planning. Both work, the robot travel path from the starting point can be better planned to a destination.
A * D * and path planning algorithm introduced
In reality, robot path planning in addition to considering the environment known and unknown environmental map, but also consider the path planning in dynamic and static environment.
A * (A-Star) algorithm is an effective algorithm for static road network in the shortest path most effective method of direct search, the search also solve many problems. Distance estimate and the actual value is closer to the algorithm, the final search faster. However, A * algorithm also can be used for dynamic route planning them, but when the environment changes, the need to re-plan the route.
The D * algorithm is a dynamic path heuristic search algorithm, which prior environmental position, allow the robot to move freely in an unfamiliar environment, its capability in a rapidly changing environment. D * algorithm biggest advantage is no need to pre proved map, the robot can and people, even in an unknown environment, you can also initiate action, continue to explore with the robot, the path will be time to adjust.
In summary, the mobile robot path planning technology has made considerable achievements, however, there are still many shortcomings in its global and local path planning method, for which the country has improved for this type of algorithm, such as thinking Lan technology SLAMWARE modular autonomous navigation, path planning algorithm using a modified D * within SLAMWARE, the core of which is the US Mars probe using pathfinding algorithm. Is a dynamic path heuristic search algorithm, which allows the robot to walk freely in an unknown environment, its capability in the case of the environment variable.