I-Corps: Smart Street Parking Assistant

这个I-Corps项目更广泛的影响/商业潜力是智能街道停车技术的开发。据估计，美国城市平均30%的交通是由寻找停车位的车辆组成的。在大城市寻找路边停车位可能具有挑战性;搜索时间很长，标志可能令人困惑或误解，错误可能是昂贵的（罚单或拖车）。这可以通过以下方式缓解：（i）街道停车规划，在出行前确定合适的停车区域，以及（ii）路边冲浪验证停车规则。智能停车场的估计市场规模为38 B美元。 该技术通过应用程序和API为驾驶员、交付提供商和智能车辆/智能城市基础设施提供这两种基本的街道停车服务。该I-Corps项目基于一系列机器学习算法和物联网技术的开发。具体地，开发了新技术并在以下步骤中使用：（i）从智能手机/车辆上的摄像头拍摄的图片/视频中检测街道停车标志;（ii）从检测到的街道停车标志中进行文本识别;（iii）用于生成停车规则的自然语言处理（NLP）;（iv）允许的停车时间计算;以及（v）众包中的信任评估。该技术主要用于街道停车标志的识别和解释。虽然有许多基于人工智能的对象检测和文本识别技术，但通常它们在这些应用中效果不佳，因为街道停车标志上的文本是不同的，并且城市之间的标志也各不相同。此外，该技术还解决了区分停车标志和其他共址标志的难题。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 I-Corps project is the development of technology for smart street parking. An estimated thirty percent of all traffic in the average U.S. city consists of vehicles searching for parking spots. Finding street parking in big cities can be challenging; the time spent searching is long, signage may be confusing or misinterpreted, and errors can be expensive (ticketed or towed). This could be mitigated with: (i) Street parking planning identifying suitable parking areas in advance of a trip, and (ii) curb-surfing verify the parking rules. The estimated market size for smart parking is $3.8 B. The proposed technology offers these two essential street parking services to drivers, delivery providers, and smart vehicle/smart city infrastructure through apps and APIs.This I-Corps project is based on the development of a series of machine learning algorithms and IoT technologies. Specifically, the novel techniques are developed and used in the following steps: (i) Street parking sign detection from pictures/videos taken by cameras on smartphones/vehicles; (ii) text recognition from the detected street parking signs; (iii) Natural language processing (NLP) used to generate parking rules; (iv) allowed parking time calculation; and (v) trust evaluation in crowdsourcing. This technology focuses on street parking sign recognition and interpretation. Although there are many AI-based object detection and text recognition techniques, typically they do not work well in these applications as the text on street parking signs is different and signs vary among cities. In addition, the technology addresses the challenge of distinguishing parking signs from other co-located signs.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.