Study On Detection And Fusion Of High-speed Train Running Information

As the basic of commands generating and decision-making for train control system, thehigh-speed train running information mainly includes train basic parameter, speed, position,movement authority, line data and external environment. In order to enhance the controlaccuracy of on-board equipment, the existing train control system should use data fusiontechnologies to synthesize different channel data to insure the information more accurate,complete and reliable which the train control unit received.As the objects of the research are the issues of train running information detection andfusion, taking Chinese train control system at level3(CTCS-3) for example, the overall fusionmodel of train control system information has been put forward. In order to get more accuratedata, the fusion schemes of train speed detection, movement authority, temporary speedrestriction and line data, which belong to on-board equipment layer, have been highlighted inthe thesis. With the different way of modeling and simulation, the feasibility and validity ofschemes have been verified.By analyzing the principle of train speed detection and configuration of speed measuringdevice, the measuring information fusion system of multiple speed sensor has been modeledby the federal kalman filter algorithm. Meanwhile, in order to real-time revise and estimatethe noise statistical characteristics, the adaptive kalman filter algorithm has been introduced torestrain filter divergence and improve fusion performance. Results show that the adaptivefederal kalman filtering algorithm has better effect than classical kalman filtering algorithm,because it can further reduce measurement errors, and obtain more stable and accurate trainspeed value. For further reliability and accuracy, the threshold has been given by analyzed thedetection and correction methods of train idle and slip. And the improvement measure ofspeed measurement also has been put forward.Meanwhile, movement authority, line data, temporary speed restriction which from thechannels of CTCS-2and CTCS-3, have been modeled according to the fusion schemes ofon-board equipment layer with CPN Tools that analyses colored petri nets(CPN). To prove theavailability of fusion schemes, the fault injection test method has been brought in to simulatethe transmission errors. The numbers of error, which occurred before and after fusion, provedthe fusion schemes can effectively reduce the output errors. These schemes are feasibletechnology way to improvement the reliability of train control system.