Event detection and classification for connected car

联网汽车的事件检测和分类

• 批准号: 519938-2017
• 负责人: Bahreyni, Behraad
• 金额: $ 0.91万
• 依托单位: Simon Fraser University
• 依托单位国家: 加拿大
• 项目类别: Engage Plus Grants Program
• 财政年份: 2017
• 资助国家: 加拿大
• 起止时间: 2017-01-01 至 2018-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=640418
• 关键词: Event; detection; classification; connected; car

Moj.io Inc. is a leader in developing backbone technologies for the rapidly growing connected car industry.Moj.io relies on information collected by a telematics dongle that is connected to the on-board diagnostic portof a car. In addition to the information provided by the engine control unit, the telematics module uses a smallnumber of sensors to provide awareness regarding the car ambient and event detection. Moj.io has beenseeking solutions to detect the glass-breakage event relying only on the limited sources of information. Theirinitial attempts on a simple threshold-based algorithm led to frequent false-positive alarms. Throughcollaborations with the SFU researchers, a modified algorithm was developed that relying on only the dataavailable from accelerometers, the glass breakage event could be reliably detected and distinguished fromsimilar events.The proposed research will build on our earlier results to (i) Implement the developed algorithm on the donglesused by Moj.io; (ii) Develop signal collection, feature extraction, and transmission algorithms at the donglelevel to minimize communication requirements; and (iii) Implement cloud-level event detection algorithmswhich to consider differences between dongle location and orientation in different cars.If successful, this will form the basis for manufacturing of future generations of telematics dongles for Moj.io,with significant commercial importance.

Moj.io Inc. 是为快速发展的互联汽车行业开发骨干技术的领导者。Moj.io 依赖于连接到汽车车载诊断端口的远程信息处理加密狗收集的信息。除了发动机控制单元提供的信息外，远程信息处理模块还使用少量传感器来提供有关汽车环境和事件检测的感知。 Moj.io 一直在寻求仅依靠有限信息源来检测玻璃破碎事件的解决方案。他们最初尝试使用简单的基于阈值的算法，但导致频繁出现误报。通过与 SFU 研究人员的合作，开发了一种修改后的算法，仅依靠加速度计提供的数据，就可以可靠地检测到玻璃破碎事件并将其与类似事件区分开来。拟议的研究将建立在我们早期成果的基础上：(i) 在 Moj.io 使用的加密狗上实施开发的算法； (ii) 开发加密狗级别的信号收集、特征提取和传输算法，以最大限度地减少通信要求； (iii) 实施云级事件检测算法，以考虑不同汽车中加密狗位置和方向的差异。如果成功，这将为 Moj.io 制造未来几代远程信息处理加密狗奠定基础，具有重大的商业重要性。