Mobile Robot Target Detection And Target Recognition Fusion Technology Research

With the rapid development and popularity of computer technology, the robotics technology has become an interdisciplinary integrated information processing technology. At present, uncertainty analysis ability of mobile robots has become an important factor to measure the robot intelligence. This article comes from the National 863 Program "service robots", key project "Home Assistant Robot intelligent integrated platform agility" and "863" plan thematic topic "in-patient-oriented intelligent space technology and patrol robot inspection system," to carry out information fusion technology directed at mobile robot. This thesis which centers on information fusion starts from uncertainty information analyzing ways to explore the mobile robot's target detection and target identification methods. The thesis research includes the following aspects:Firstly, this paper introduces the research background and the development of target recognition and mobile robots' information fusion at home and abroad, meanwhile, different definitions of information fusion will be introduced.Secondly, this uncertainty information analysis methods are outlined, the causes of uncertainty information in the work of mobile robots are analyzed in this paper. Besides, the main content of this research is recommended.Thirdly, to the uncertainty information in the work of mobile robots, the electro circuit design of ultrasonic ranging sensors is carried out in the process of the target detection of mobile robots, while the results of ultrasonic ranging data was calibrated, the ranging linear regression model of ultrasonic sensor is given through the mathematical derivation and a hypothetical obvious test is concluded; the single application of ultrasonic sensors can not meet the robots' precise identification to obstacles, the combined use of other sensors to improve the measurement accuracy to the obstacles becomes necessary. The paper has hardware and software design to the ranging distance of infrared sensors and the PC communication program design of the infrared sensor is derived. Fourthly, after detecting the target, the paper analyzes the overview of mobile robot target recognition information fusion of multi-sensor as well as the existing difficulties in mobile robots' target recognition. This paper extends the conceptions of relevant coefficients into the robots' target recognition and establishes the similarity coefficient mathematic models of robot target recognition. At the same time, a target recognition fusion algorithm based on maximum a posteriori probability is put forward and experimental verification of the algorithm is carried out so that the fact that the algorithm can improve the identification rate and reliability of the system can be proved.Fifthly, to the robot object recognition, this paper proposes a structural method of basic probality function in D-S evidence theory by using D-S evidence theory to carry out target recognition fusion. This method plays a significant role in the decisive result.Sixthly, this paper improves the application of D-S theory in target recognition fusion and puts forwards the thought of two-level fusion of time and space. Combined with D-S evidence theory, the fusion algorithm to the data of time and space data is derived, which greatly improves the result in mobile robots' target recognition. As a result, the mobile robot system is equipped with very good competence to Fault-tolerant. Meanwhile, the paper also proposes the target fusion algorithm based on the fuzzy integrated algorithm.Finally, this paper sums up research results, puts forward the existing problems in the research and expounds the work still needed to be done in the field.