Apollo低成本方案在TX2平台针对相对地图的优化

Optimization for Relative Map of Apollo Low-Cost Solution on TX2 Platform

自动驾驶技术是一项正在逐步改变人类出行方式的高科技技术。Apollo是百度公司推出的自动驾驶开源平台,已支持L3级别的限定区域的自动驾驶和L4级别的城市复杂路况的高度自动驾驶。其中相对地图方案以低成本,高可靠性在L3级的封闭园区及高速路方案中得到广泛应用。在不同的场景中,相对地图方案的路线长度不同。当路线过长时会因为加载和刷新路线数据而导致计算性能下降。针对该问题,设计一种相对地图的局部加载方案,以一种可控且不影响车辆运行状态的算法达到节省计算资源的目的。

Autonomous driving technology is a developing high technology that is changing the way of people traveling.Apollo is an open-source au.tonomous driving platform launched by Baidu Corporation,now it supports geo-fenced autonomous driving with the level 3 and advanced autonomous driving in complex downtown areas with level 4.Relative Map is used widely in closed venue low-speed and highway solutions because it’s highly reliable and low-cost.The lengths of routes are different in different environments.If the route is too long,loading and updating large route data could cause computing performance to decrease.To solve this issue,designs a new partially loading solution for Relative Map,aims to save computing resource and make vehicles drive more safely and smoothly.