Investigating the use of MEMS Based Wrist-worn IMU for Pedestrian Navigation Application

J. Qian, J. Ma, L. Xu, R. Ying, W. Yu, P. Liu

Abstract:	As one of the most challenging applications in navigation technologies development, pedestrian navigation has gained great concern in recent years. Pedestrian navigation system (PNS) is generally required to provide continuous positioning capability in all environments including urban canyons, indoors and undergrounds, where GPS and other radio navigation signals may degrade or even outage, in terms of accuracy and availability. With the development and evolution of micro electromechanical system (MEMS) technologies, it becomes possible to integrate inertial measurement units (IMUs) into personal navigation devices. Also the IMU devices have several advantages over the ubiquitous GPS system, such as small size, light weight and especially low power consumption, which are the major concerns for pedestrian navigation applications and make low-cost MEMS based IMUs regularly utilized as the augmentation of various pedestrian navigation solutions. However, MEMS based IMUs are only able to provide required accuracy (the positioning error must be less than 5 meters in many situations, such as indoor positioning) for brief moments due to the sensor errors arise from random zero bias and oscillation noise. Moreover, the errors correction and compensation using GNSS or other position systems are restricted in the low dynamic and high vibration environment. As we know, introducing zero velocity updates (ZUPTs) as pseudo measurements into the Extended Kalman Filter (EKF) is an effective way to reduce the error drift of an inertial sensor-based navigation system. The main concept consists in detecting the stationary stance phase during pedestrian walking cycles and resetting the velocity error to zero, consequently making the position and velocity errors diverge slowly and improving the navigation solution performance. However, the main drawback of this methodology related to inconvenience has been discussed by several researchers. They point out that the IMU module, in the ZUPT algorithm, has to be attached on the foot in all the experiments, which may bring some inconvenient and even uncomfortable feelings for the pedestrian. In addition to this method, pedestrian dead reckoning (PDR) also can provide means of reducing the inertial error accumulation to the navigation solution by taking advantage of the sequential characteristics of the pedestrian motion. The algorithm is a relative navigation technique, which determines the relative location of a pedestrian by using step detection, stride length estimation, and heading determination. Typically, the accelerometer measurements are utilized to implement step detection and stride length estimation, and heading determination is simultaneously completed by fusing the information from gyroscopes and magnetometers. This paper investigates a novel scheme using MEMS based wrist-worn IMU for pedestrian navigation application. The mounting of the IMU sensors on the wrist is, compared with other parts of the human body (such as foot, waist, and hip), more convenient but meanwhile more complex due to random activities of hands. Therefore, the performance of previous algorithms may be unable to be guaranteed when the measurements are collected from wrist-worn IMU. On the other hand, the flexibility of hands also provides the opportunity for us to recognize and deduce the behaviors of pedestrians in the experiments. First of all, this paper focuses on the topic of the designing of a robust step detection algorithm based on wrist-worn accelerometer data. The algorithm depends on the fact that the total acceleration magnitude exhibits cycles typical of a human’s walking motion. Using the local gravity value crossings detection and autocorrelation operation of measured acceleration signals, we detect steps with low false alarm probability, which makes sure that the algorithm is not susceptible to the actual situation and independent of walking patterns, routes, distances and terrains. Then, the stride length is estimated based on the relationship between stride length and stride interval and the recognition and classification of pedestrian walking patterns. Extensive experiments prove the fact that the precision of the stride length estimation is not only various among different pedestrians, and influenced by many factors, such as the slope of the terrain, the walking activities, the motion modes, and so on. Thus the walking activities recognition and motion modes classification, including stationary and walking, forward and backward, fast walking and running, stairs and elevators, are performed first based on wrist-worn IMU measurements. Hereafter, stride length estimation models are established according to the classification results and are tested repeatedly in order to find out the most appropriate parameters for the pedestrians. Finally, in this paper we present a new technique to correct gyroscopes drift due to operation over longer durations of time by making use of the periodicity of swing arms during walking. This technique extends the period in which the relative heading information from gyroscopes remains quite reliable. Moreover, the accuracy of the result is further improved by integrating gyroscopes and magnetometers in the EKF design. The experiment results show that the scheme proposed in this paper can achieve superior performance in terms of positioning accuracy, reliable operation time, power consumption, and convenience of application. As described above, the proposed scheme base on wrist-worn IMU module is worthy of further exploration for pedestrian navigation application. It can be concluded that through using appropriate physiological models and advanced algorithm, it would be possible to implement a pedestrian navigator using available MEMS based wrist-worn IMU sensor technology.
Published in:	Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013) September 16 - 20, 2013 Nashville Convention Center, Nashville, Tennessee Nashville, TN
Pages:	1057 - 1064
Cite this article:	Qian, J., Ma, J., Xu, L., Ying, R., Yu, W., Liu, P., "Investigating the use of MEMS Based Wrist-worn IMU for Pedestrian Navigation Application," Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013), Nashville, TN, September 2013, pp. 1057-1064.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In