The Environment Map Building Based On Laser Sensor

With the development of economy nowadays,private cars become more and more popular.But people worry about the traffic problems caused by the growing numbers of vehicles while enjoying the convenience.The traffic accidents not only cause economic damage to the drivers,but also threaten their lives.In order to reduce the number of traffic accidents,people pay more and more attention to the research on automatic driving technology.The primary work of automatic driving technology is to detect the obstacle in front of the vehicle and then generate the corresponding environment map.Although the research on the construction technology of the environment map has already achieved fruitful results,there are still some challenges in obstacle detection,real-time update and so on.In this paper,the laser sensor is used to obtain the information of obstacles in front of the vehicle,and the information is collected and preprocessed.Then a conclusion of obstacle detection is drawn by the data analysis.Firstly,Describe the key problems of environment map building and the basic knowledge of the multi-sensor information fusion,which lay the theoretical foundation for the experiment of the building of the environment map based on laser sensor.Then,for the problem of building the environment map,the paper introduces the method of probabilistic grid to study the problem of setting up the optimal grid resolution,and reduced the computational quantity while ensuring the accuracy of the experiment.For real-time update problem,the modified Bayesian method is used for the data fusion of different moments to solve the obstacle occlusion problem in a single moment.The experimental results in simulation and real environment showed that,the modified Bayesian method realized the map update effectively.For the problem of obstacle detection,this paper proposes the improved method of the image of average and variance,and compares with the result of obstacle detection by occupied or blank method.Experimental results showed this method improved the accuracy of detection,and solved the problem of the obstacle detection effectively.Finally,a single laser sensor can only collect one frame data per moment,and the range of detection is limited,and there are some errors in the measurement.Aimed at this problem,four laser sensors are used to collect information in this paper.Dempster-Shafer theory of evidence is used in information fusion.Then the software player/stage is used to scenario simulation.Experimental results showed that four laser sensor data fusion works effectively by the Dempster-Shafer theory of evidence.The errors of single sensor environment map were reduced.And the construction of a global map of the surrounding of vehicles was achieved.