Research On Detection Method Of Obstacle In Front Of Train Operation Based On Radar Measurement Data

Railway transport as one of the three main modes of transport,passenger travel and the flow of bulk cargo transport plays an important mission.China's railways have long mileages and complex operating environments.Frequent incidents caused by foreign objects invading railway line building boundaries constitute a serious threat to the safe operation of trains.When the train is running at a high speed,it is no longer possible to meet the safe operation of the train by relying on the traditional driver's visual way of driving.Therefore,it is necessary to use the automatic monitoring equipment to assist the driver in real-time monitoring and perception of the running environment ahead,that is,to gradually realize the full-automatic driving of the train.This thesis analyzes the research status and shortcomings of foreign object intrusion detection at home and abroad,and proposes the use of radar sensors to detect foreign object intrusion detection on railway lines.The main research contents are as follows:(1)The segmentation fitting of the measured trajectory data obtained by GPS is obtained to obtain the orbital map piecewise polynomial expression for auxiliary positioning and construction of the detection area.Firstly,the track line data point information obtained by GPS is processed through coordinate transformation to obtain track line data points based on the engineering coordinate system.Secondly,the track line data points based on the engineering coordinate system obtained after the transformation are used to eliminate the gross error data in the data point set by using the angle discrimination method to obtain the initialized track line data points.Thirdly,The scatter diagram of the slope change between adjacent points of the track data after initialization is analyzed,and the shape of the track line is determined,and it is divided into segment track data points.Finally,the piecewise polynomial expressions of the orbit map obtained by piecewise fitting are used to assist vehicle radar positioning and detection area construction.(2)The detection area model is constructed by combining the positioning information of the vehicle-mounted radar provided by GPS,the segmented polynomial expression of the orbit map,and the design specifications of the railway line building limits.Firstly,the radar position information provided by GPS is acquired in real time when the train is running.Secondly,the radar position information is matched to the segmented polynomial expression of the orbit map to predict the coordinate position of the radar transmitting point in the orbit map.Finally,the detection area model is constructed by the least square method based on the coordinate position information of the radar transmitting point in the orbit map,the parameter information provided by the radar,and the segmental polynomial expression of the orbit map.(3)The target position information acquired by the radar is preprocessed and substituted into the constructed detection area model for obstacle detection and judgment.Firstly,the pre-processing link is used to implement the error processing on the original target position information acquired by the radar.Secondly,the position information of the target point based on the radar coordinate system is transformed into the engineering coordinate system by plane coordinate transformation.Finally,the converted target point position information is substituted into the detection area model to detect obstacles and output the detection results.The field test results show that the straight rail detection accuracy rate is 90.50%,the curved rail detection accuracy rate is 86.79%,the turnout area detection accuracy rate is 82.67%,and the average accuracy rate is 87.45%,which have good detection performance.