SBIR Phase I: Risk-Aware Motion Planning for Autonomous Vehicles

该项目更广泛的影响/商业潜力是使自动驾驶汽车能够高速做出风险意识决策，通过推理周围人和汽车的不确定运动来预测风险并做出防御反应。这项技术将首次使安全和风险的新技术方法变得可行。它将使自动驾驶汽车能够在人口密集的城市环境中以正常速度行驶，同时降低事故发生的可能性-从而实现更快，更安全的城市自动驾驶。这将提高公众对自动驾驶汽车的信任度，并加速一个新行业的发展，该行业将重塑交通运输，使人们不再需要驾驶自己的汽车，并增加他们的空闲时间和生产力。这一小型企业创新研究（SBIR）第一阶段项目旨在开发一种专用处理器，用于自动驾驶汽车的风险感知运动规划。运动规划是自动汽车的一个关键问题，它必须决定如何在空间中移动以达到指定的目标，而不会与静态障碍物、另一辆车或行人相撞。现有的方法只能每秒构建几个计划，并假设其他道路使用者将继续他们当前的轨迹。这在高速公路上是可以接受的，但在城市驾驶中就不行了，因为城市距离很短，其他道路使用者可能会突然改变行为。该项目旨在开发一个专门的运动规划处理器的原型，该处理器可以在毫秒内进行规划，使其能够考虑其他道路使用者未来所有可能的运动，以预测风险并做出防御反应。该提案旨在通过设计原型来评估处理器的可行性，证明其在城市环境中以正常速度行驶时的安全性，并量化其相对于现有解决方案的性能改进。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this project is to enable autonomous vehicles to make risk-aware decisions at high-speed, by reasoning about the uncertain movements of the people and cars around them to anticipate risks and react defensively. This technology will make new technical approaches to safety and risk viable for the first time. It will enable autonomous cars to drive at regular speeds in dense, highly populated urban environments while reducing the likelihood of accidents - thereby resulting in faster, safer urban autonomous driving. This will increase public trust in autonomous cars and accelerate the development of a new industry which could reshape transportation, freeing people from the need to drive their own cars and increasing their free time and productivity.This Small Business Innovation Research (SBIR) Phase I project aims to develop a specialized processor for risk-aware motion planning for autonomous vehicles. Motion planning is a critical problem for autonomous cars, which must decide how to move through space to reach a designated goal without colliding with a static obstacle, another vehicle, or a pedestrian. Existing approaches can only construct a few plans per second, and assume that the other road-users will continue on their current trajectories. This is acceptable for highways but not for urban driving, where distances are short and other road-users can change behavior suddenly. This project aims to develop a prototype of a specialized motion planning processor that plans in milliseconds, enabling it to consider all likely future movements of other road-users to anticipate risks and react defensively. This proposal aims to evaluate the feasibility of the processor by designing a prototype, demonstrating that it results in safer driving at normal speeds in urban environments, and quantifying its performance improvements over existing solutions.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.