Font Size: a A A

Research On Utility Optimization For Outdoor Mobile Robot

Posted on:2016-04-04Degree:DoctorType:Dissertation
Country:ChinaCandidate:F LiuFull Text:PDF
GTID:1108330479983249Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of computer, control theory and artificial intelligence, it shows increasing requirement in application for robots in military, industry, agriculture and domestic services etc. The workspace extends from simple indoor environment to the complicated outdoor environment such as space, underwater environment and special terrains. There has a particular type of robot that is widely utilized in large-scale outdoor environment, e.g., the transformer substation inspecting robot, the auto-guided vehicle for carrying cargo and the reconnaissance robot. They need to work continuously and long-termly, and will be frequently ordered to visit one or multiple goals to execute special tasks. For these mobile robots, to optimize the utility is the key objective in practical application, and is what this dissertation studies.Focusing on the utility of the robot, this dissertation explores the essential factor that can reflect the optimal utility and studies how to optimize the utility in real application. Reducing energy consumption can increase the time for working effectively and thus is an important way to improve the utility. Thus, a new energy consumption model is erected based on considering multiple road attributes in outdoor environment. Based on the model and the characteristics of this application, idle time model is proposed that can reflect the utility in essence. Afterwards, aiming on the multi-goal visiting task, the method for optimizing utility is studied, in which optimal path planning and multi-task allocation are used for a single robot and multi-robot system. The main work of this dissertation includes four parts:①To establish the energy consumption model on consideration of multiple road attributes for outdoor environment. Currently, the analysis on energy consumption is mainly aiming on indoor environment, while for outdoor environment only the energy consumption on movement is discussed. For outdoor environment, it is more complicated as it has special attributes such as road surface roughness and road grade. Taking into account path length, roughness and road grade, the energy consumption is analyzed in detail in which the part on movement is greatly discussed, and finally the energy consumption model is obtained.②Based on the developped energy consumption model, idle time model is further proposed which is the substantial factor reflecting utility. For the robot working continuously and long-termly, to maximize utility means that the robot has the maximum working time. Oppositely idle time(non-working time) is proposed to describe the essential factor that represents utility. Idle time includes charging time and maintenance time. The computation of maintenance time is obtained by analyzing the main issue causing fault, which is the continuous vibration induced by multiple road attributes. Lastly, with charging time, the computation of idle time is achieved. Therefore, utility maximization is to minimize the idle time.③By use of energy consumption model and idle time model, a tailored genetic algorithm is proposed to plan optimal path for the robot visiting multiple goals in order to achieve the maximum utility. For multi-goal visiting problem, the robot needs to visit the goals with minimum cost — idle time while under the constraint of limited remaining energy. Thus, based on the task’s characteristics, modification is made on the basic genetic algorithm to plan the optimal sequence for visiting goals as well as the optimal path for robot. Consequently, the effectiveness is verified by simulations.④For the problem that multiple robots visiting multiple goals to execute particular tasks, a system utility model is presented, and for obtaining the optimal utility, multi-robot task allocation is employed using the proposed distributed-computing and centralized determination(DCCD) method. A utility model is first developed by considering the cost of completing task(idle time) and the quality of task accomplishment. Then, DCCD is proposed which comprises the advantage of centralized and distributed task allocation methods. At last, utility model and DCCD are employed to allocate tasks to robots in order to maximize the system’s utility. The effectiveness and superiority are proven by simulation.
Keywords/Search Tags:Mobile Robot, Utility, Idle Time, Optimal Path Planning, Multi-Robot Task Allocation, Genetic Algorithm
PDF Full Text Request
Related items