Robot Navigation, Perception and Planning for Intelligent Energy Management in Electric Vehicles

据独立预测，到2020年，英国道路上将出现数十万辆插电式电动汽车和混合动力汽车。虽然这项技术的采用目前是受到环境问题的推动，但电动和混合动力汽车技术对可持续经济增长的巨大潜力正变得越来越明显。然而，为了让这项技术达到所需的渗透率，成为传统汽车的可行大众市场替代品，它需要被视为满足了消费者的需求。最近的研究表明，由于对续航里程限制和车辆成本的担忧，纯电动汽车的这种大众市场渗透主要受到阻碍。车辆对剩余可用里程的不准确反馈加剧了“射程焦虑”。成本的上升主要是由于电池容量和寿命的限制，这两者都受到充电周期的数量和激烈程度的影响。众所周知，电动汽车的续航里程以及最终的充电需求受到许多因素的显著影响，例如车速分布和车辆轨迹上的地理位置、路况或天气。例如，在交通拥堵的情况下反复加速上山，比在平地上以恒定的速度巡航更消耗负荷。然而，这些见解几乎没有改善个人最终用户的体验：目前在日常部署中考虑车辆续航里程或电池寿命时，既没有利用驾驶员特定的信息，如驾驶行为或常用驾驶路线，也没有利用特定路线的信息，如交通量、速度限制或停车标志和红绿灯的位置。该项目通过利用最先进的机器人学和机器学习技术来预测车辆行驶里程并优化电池寿命，从而解决了这些缺点。在机器人导航和感知的背景下建立的方法非常适合于提供关于驾驶员行为和路线基础设施的不断演变的现场信息。与这种特定于驾驶员的汽车使用概况相结合，涉及稳健规划和决策的核心机器人技术可以解决决定何时充电以及充电多长时间的任务，从而保护电池寿命并将充电成本降至最低。因此，通过考虑车辆的行驶方式、地点和时间，该项目将改进对车辆行驶里程的预测，并产生更贴近实际的收费制度。

By 2020 independent forecasts predict hundreds of thousands of plug-in electric and hybrid vehicles on UK roads. While the adoption of this technology is currently driven by environmental concerns, the significant potential of electric and hybrid vehicle technology for sustainable economic growth is becoming increasingly apparent. However, in order for this technology to achieve the penetration required to become a viable mass-market alternative to conventional cars it needs to be perceived as meeting consumers' needs. Recent studies have shown that this mass-market penetration is primarily impeded for all-electric vehicles by fears over range limitations as well as vehicle cost. 'Range Anxiety' is fueled by inaccurate feedback from the vehicle regarding the remaining range available. Costs are driven up primarily by limitations on battery capacity and life, both of which are affected by the number and ferocity of charging cycles.It is an established fact that the range of an electric vehicle, and therefore the eventual need for charging, is significantly influenced by a number of factors such as the velocity profile and geography along the vehicle's trajectory, the condition of the road or the weather. Repeatedly accelerating up a hill in a traffic jam, for example, is more load-intensive than cruising at constant speed on level ground. However, few of these insights improve the experience of the individual end-user: neither driver-specific information such as driving behaviour or commonly driven routes nor route-specific information such as traffic volume, speed limits or the location of stop-signs and traffic lights are currently exploited when considering vehicle range or battery longevity in every-day deployment. This project addresses these shortcomings by leveraging state-of-the-art Robotics and Machine Learning techniques for the prediction of vehicle range as well as the optimisation of battery longevity. Methods established in the context of robot navigation and perception are ideally suited to provide evolving, in-situ information on driver behaviour and route infrastructure. In concert with such a driver-specific usage profile of a car, core robotics technologies concerning robust planning and decision making can address the task of deciding when and how long for to charge a vehicle such that battery life is preserved and charging costs are minimised. Therefore, by considering how, where and when a vehicle is traveling this project will lead to improved forecasts of vehicle range as well as to more germane charging regimes.