Route-planning based on a passenger condition for self-driving vehicles
Route-planning based on a passenger condition for self-driving vehicles
复制标题
DOI：
10.1109/icci-cc.2017.8109769
复制
发表时间：
2017-07
期刊：
影响因子：
--
通讯作者：
H. Hiraishi
中科院分区：
文献类型：
--
作者：
H. Hiraishi
本文提出了一种针对自动驾驶车辆广泛使用的环境的路线规划方法。此类车辆将根据乘客的状况规划路线，并使用生物传感器识别乘客的状况。多家汽车制造商和 IT 公司最近开发了各种自动驾驶汽车技术，谷歌公司就是其中之一。在本文中，我们重点关注路线规划技术。我们特别讨论了如何实现避免交通拥堵。我们的自动驾驶车辆会生成一条新路线，以避免交通拥堵。我们采用了时间约束启发式搜索（TCS），我们可以提前设置时间限制。如果我们设置更长的时间限制，则可以生成更接近最优路线的路线。 TCS确保车辆在进入交通拥堵之前无需停车即可获得避让路线。我们使用我们自己开发的交通模拟器对自动驾驶车辆的总效率和扩散率之间的关系进行了一些实验。结果，我们能够阐明总效率出现峰值并随着更多车辆产生回避路线而下降的现象。因此，生成回避路线并不总是最好的，并且沿着当前路线行驶而不生成回避路线的决定在某些情况下变得重要。因此，我们提出了一种车辆根据乘客状况判断是否生成避让路线的方法。为了检测乘客的状况，我们使用了坐压传感器，它可以检测乘客重心的移动情况。该传感器使我们能够成功识别乘客的疲劳程度。因此，我们可以做出一定的判断：如果乘客看起来很放松、心情舒适，车辆会沿着当前路线行驶；如果乘客看起来很疲倦或烦躁，车辆会提前到达，避免交通拥堵；如果乘客看起来明显疲倦，车辆会停下来休息一段时间。
This paper proposes a route-planning method for an environment in which self-driving vehicles are widely used. Such vehicles will plan their route based on the passenger's condition, which is recognized using a biological sensor. Various carmakers and IT companies have recently developed various technologies for self-driving vehicles, Google, Inc. being one. In this paper, we focus on a technique for route planning. In particular, we discuss how to realize the avoidance of traffic congestion. Our self-driving vehicles generate a new route to avoid traffic congestion when it occurs. We adopted a time-constrained heuristic search (TCS) to which we can set the time limit in advance. If we set a longer time limit, routes closer to the optimal route can be generated. A TCS ensures that vehicles can obtain an avoidance route without stopping before entering the traffic congestion. We executed some experiments concerning the relationship between total efficiency and the diffusion rate of self-driving vehicles using our own self-developed traffic simulator. As a result, we were able to clarify the phenomenon in which a peak in the total efficiency occurs, and decreases as more vehicles generate avoidance routes. Therefore, it is not always best to generate an avoidance route, and the decision to drive along the current route without generating an avoidance route becomes important in certain cases. Thus, we propose a method in which a vehicle judges whether to generate an avoidance route based on the passenger's condition. To detect the passenger's condition, we use a sitting-pressure sensor, which can detect the movements of the passenger's center of gravity. This sensor allows us to succeed in recognizing passenger fatigue. We can therefore make certain judgments: The vehicle will go along the current route if the passenger seems to be relaxed and in a comfortable atmosphere, the vehicle will arrive earlier by avoiding traffic congestion if the passenger seems to be tired or irritated, or the vehicle will stop for a break period if the passenger seems to be significantly tired.