PRE CHASM RESEARCH LTD - MyTyreManager - a Mass Market Smart Phone Tyre Tread Technology Using Novel Active Machine Learning

该项目旨在让机器代替人对轮胎磨损做出复杂的决定，建立在成功完成的TSB ICT技术可行性项目（FileRef 131363）的基础上。该可行性研究表明，目前的3D扫描仪价格昂贵（> 1000英镑），不适合执法人员或消费者（车主/司机）使用。我们发现，手动方法容易被误解，往往被忽视。我们证明了将智能机器学习和新算法嵌入智能手机进行轮胎磨损图像分析的可能性，首次允许基于智能手机的“机器”对车辆轮胎的道路适用性做出复杂的决定。该项目旨在将该技术发展到概念验证阶段，目标是开发消费者“应用程序”和平台架构，使车主能够使用智能手机持续管理轮胎磨损和状况。这将使用户能够改变他们的驾驶风格，以优化轮胎寿命，并最大限度地提高安全性。（如保险公司）可以分析这些数据。需要解决的挑战包括，在一系列智能手机上优化特征向量算法，轮胎差异和对机器视觉的影响，学习和准确性，实时云集成，天气的影响，以及优化采集速度。2011-2020年是联合国道路安全行动十年，全球道路交通事故造成130万人死亡。增长30%/年。45%与轮胎磨损有关。根据轮胎安全组织的数据，仅在英国就有440万辆汽车至少有一个非法轮胎。问题是巨大的。因此，服务市场同样巨大。其中一个指标可能是全球轮胎销量与智能手机用户的对比，这表明一个超过5亿用户的可服务市场。这是一个在消费者“应用程序”层面广泛采用专利机器学习的机会，也是一个对道路安全产生积极影响的机会。

This project aims to allow machines instead of people to make complex decisions abouttyrewear, building on a successfully completed TSB ICT Technical Feasibility project (FileRef 131363).That feasibility study stated that current 3d scanners are expensive (>£1000) and not fit foruse by law enforcers, or consumers (car owner/drivers). We showed that manual methods areopen to misinterpretation and are often ignored. We demonstrated that it was possible toembed intelligent machine learning and novel algorithms into a Smart Phone for tyre-wearimage analysis, allowing for the first time, a Smart Phone-based 'machine' to make complexdecisions about the road-worthiness of tyres on vehicles.This project seeks to progress that technology to Proof of Concept stage, targeting thedevelopment of a consumer 'App' and platform architecture to enable car owners to managetheir tyre wear, and condition, on an on-going basis using their Smart Phone. This will enableusers to alter their driving style to optimise tyre longevity, and maximise safety.Other interested parties (like insurers) could analyse that data.Challenges to address include, optimisation of feature vector algorithms across a range ofSmart-Phones, characterise tyre differences and impact on machine vision, learning andaccuracy, realtime cloud integration, influence of weather, and optimisation of acquisitionspeed.Significance: 2011-2020 is the United Nations Decade of Action for Road Safety(UN).Worldwide road accidents kill 1.3million. Increasing 30%/yr. 45% linked to tyre wear.According to TyreSafe in UK alone 4.4m cars have at least 1 illegal tyre. The problem isenormous. The serviceable market is equally large therefore. One measure could be tyres soldvs smart phone users worldwide, suggesting a serviceable market in excess of 500m users.This is an opportunity for wide adoption of patented machine learning at the consumer ‘App’level and a chance to positively impact road safety.