基于上下文感知的乘车感知算法研究和实现

发布时间：2018-11-26 19:24

【摘要】：随着乘客对交通工具乘坐舒适度的要求日渐上升,乘车舒适度已被确定为影响乘客乘车满意度和智能交通系统面临的一个必不可少的问题。在行驶过程中,驾驶员的驾驶事件和道路状况是影响乘客的乘车舒适度的主要因素。实践证明,乘车感知的研究对提高乘客的乘坐舒适度起到了积极的作用,而有效的检测出驾驶事件和每条道路上的乘车感知改善乘车感知的有效途径。目前的智能手机设备的硬件配置不断攀升,手机上的传感器越来越多,包括加速度传感器,陀螺仪,重力传感器等等,但这些传感器并没有真正做到物尽其用。本文在基于移动智能手机设备的基础上,以“参与式感知”为核心理念,引入乘客的乘坐舒适度,提出了一个基于上下文感知的乘车感知算法的研究和实现,用于检测和评估乘车感知。该系统在加速度计和磁力计的基础上,首先提出了一个坐标转换系统,通过将手机坐标系上的加速度数据通过两次转换为车辆坐标系上的加速度数据,从而消除了由手机在车上放置位置和方位的任意性给加速度传感器带来的影响,增强了用户的体验感。系统将乘车感知划分为水平方向上的乘车感知和垂直方向上的乘车感知来分别进行检查和评估。垂直方向上的乘车感知受到道路状况的影响,比如坑洞、陡坡等,这个方向上的乘车感知的评估依赖于垂直方向上的加速度数据;水平方向上的乘车感知则受到一些驾驶事件的影响,比如加速前行、刹车、转弯等等,这个方向上的乘车感知的评估依赖于水平方向上的加速度数据。将道路状况和驾驶事件作为上下文感知信息,系统设计并实现了两个评分机制,分别为道路状况评分机制和驾驶事件评分机制,给用户提供一个乘坐过程中对各类事件的直观且量化的乘车体验。同时,本文在ISO2631中对人体振动评价的基础上,提出并设计了整个行驶过程中对乘车感知的综合评价机制,更接近于人体的真实乘车感受。最后,系统根据评定的驾驶事件进一步判定出驾驶员的驾驶行为风格,并对存在危险驾驶行为的驾驶员发出安全预警信息。整个过程都是在智能终端上完成的,成本低廉,实时性强,准确度高,易于使用。
[Abstract]:With the increasing demands of passengers on the comfort of vehicles, ride comfort has been identified as an essential problem affecting passenger satisfaction and the intelligent transportation system (its). During driving, driver's driving events and road condition are the main factors affecting passenger's ride comfort. Practice has proved that the study of passenger perception plays a positive role in improving passengers' ride comfort, and effectively detects driving events and the effective way to improve the vehicle perception of each road. At present, the hardware configuration of smartphone devices is increasing, and there are more and more sensors on the phone, including accelerometers, gyroscopes, gravity sensors and so on, but these sensors are not really making the best use of things. On the basis of mobile smartphone devices, this paper introduces the passenger ride comfort and proposes a context-aware ride-aware algorithm based on the core concept of "participatory perception". Used to detect and evaluate vehicle awareness. Based on accelerometers and magnetometers, a coordinate conversion system is proposed, which converts acceleration data from mobile phone coordinate system to acceleration data in vehicle coordinate system twice. Thus eliminating the impact of the arbitrariness of mobile phone position and position on the vehicle to the acceleration sensor and enhancing the user's experience. The system divides the sense of riding into horizontal perception and vertical sense to check and evaluate respectively. The perception of the vehicle in the vertical direction is affected by the road conditions, such as holes, steep slopes and so on. The evaluation of the perception of the vehicle in this direction depends on the acceleration data in the vertical direction. The sense of the car in the horizontal direction is affected by some driving events, such as acceleration, braking, turning and so on. The evaluation of the sense of the car in this direction depends on the acceleration data in the horizontal direction. Taking road condition and driving event as context-aware information, two scoring mechanisms are designed and implemented, one is road condition scoring mechanism and the other is driving event scoring mechanism. Provide users with an intuitive and quantifiable ride experience of various events during the ride. At the same time, on the basis of the evaluation of human vibration in ISO2631, this paper proposes and designs a comprehensive evaluation mechanism for the vehicle perception in the whole driving process, which is closer to the real experience of the human body. Finally, the system further determines the driver's driving behavior style according to the evaluated driving events, and sends out the safety warning information to the driver who has dangerous driving behavior. The whole process is completed on the intelligent terminal, the cost is low, the real-time is strong, the accuracy is high, easy to use.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP301.6

【相似文献】