Enabling Affordable Autonomy Using Hybrid Dense Vision

使用混合密集视觉实现经济实惠的自主性

• 批准号: 132267
• 金额: $ 22.13万
• 依托单位: OXBOTICA LIMITED
• 依托单位国家: 英国
• 项目类别: Feasibility Studies
• 财政年份: 2016
• 资助国家: 英国
• 起止时间: 2016 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=132267
• 关键词: Enabling; Affordable; Autonomy; Using; Hybrid

Autonomous vehicles have vast economic potential for UK plc. Like any new technologies, the adoption curvecan be dramatically accelerated by improving the balance between cost and benefit. However this delicatebalance is often overlooked in many studies and sensor costs currently limit any reasonable business case.This proposal is about inducing a step change in the affordabililty of autonomous vehicles that will acceleratethis nascent technology to deployment. The Feasibility Study will combine new, ground-breaking IntellectualProperty in 3D dense vision into a road-going prototype and assess the performance of this radical low-costalternative to laser localisation over many thousands of miles of testing. The outcome will be an analysis of theperformance of a new Hybrid Dense Vision approach - leading to a reference design for an ultra-low cost sensorsystem for autonomous vehicles that could reduce existing sensors costs for autonomous systems by over anorder of magnitude.

自动驾驶汽车对英国公司来说具有巨大的经济潜力。像任何新技术一样，通过改善成本和收益之间的平衡，可以大大加快采用曲线。然而，这种微妙的平衡在许多研究中经常被忽视，传感器成本目前限制了任何合理的商业案例。这一提议是关于诱导自动驾驶汽车的可负担性的一步变化，这将加速这一新兴技术的部署。可行性研究将把联合收割机在3D密集视觉中的新的、突破性的知识产权结合到公路原型中，并在数千英里的测试中评估这种激进的低成本激光定位替代方案的性能。其结果将是对新的混合密集视觉方法的性能进行分析，从而为自动驾驶汽车的超低成本传感器系统提供参考设计，可以将自动驾驶系统的现有传感器成本降低一个数量级以上。