An Acceleration-sensor-based Trajectory Detection System Of A Moving Object

Trajectory tracking detection of moving objects has been widely applied in various fields suchas automotive brake boot detection, navigation system, human-computer interaction and gamecontrol.. Some traditional solutions, for example, magnetic space tracking system and lasertracking system, possess high measuring accuracy. However, most of them are confined by externalreference frame, narrow applicable scope, complex equipment and high cost. With the advancementof MEMS processing technology, resolution of acceleration sensor have got improved to a largeextent. Trajectory tracking detection system based on MEMS acceleration sensor features of lowcost, meeting large-scale request of business and civilian application..The principle of trajectory tracking is that three-dimensional velocity and displacement can beobtained through acceleration integration. In this paper, acceleration sensor ADXL345and ARMdevelopment board MINI2440are used to get moving objects acceleration data. In order to reducesystematic error and eliminate random noise, a variety of error processing algorithms are studied viausing the Kalman filter. Moreover, a dual speed algorithm is designed to compensate nonlinearityerror based on the random nature of the kernel process scheduling algorithm. Thus Systematicaccuracy has been significantly improved without increasing computational burden. According tothe characteristics of floating-point instruction system on ARM platform, square root and matrixmultiplication algorithms have been optimized. Hence computational performance of the system isenhanced remarkably.In this paper, all the algorithms are accomplished on ARM hardware platform and in the Linuxenvironment. Acceleration sensor in I2C subsystem and input subsystem driver are designed anddeveloped. A low coupled, scalable error compensation and trajectory restore software system isachieved in Linux practical level. Experimental results show that the system has the advantages oflow power consumption, low load and high precision, approaching the criteria for realisticapplications.