Development of the next generation of smart salt trucks for sustainable winter road maintenance

在加拿大，每年有超过500万吨的道路盐被用于道路和停车场的除冰。这些盐会对水质产生重大影响。在提供道路盐施用计划时，环境和健康与安全问题变得越来越重要。该研究项目将开发新的模型，用于规定适当数量的道路盐，并在适当的时间在给定的风暴事件中明智地应用于道路和停车场。该模型将使用遥感和地面实时天气预报数据来规定最佳的道路撒盐率和在风暴事件中施用盐的时间，以实现道路安全，同时最大限度地减少对易受盐伤害的地区的危害。我们将使用道路天气信息系统（RWIS）更准确地预报冬季风暴事件的严重程度和类型。然后，我们将使用先进的人工智能建模技术，包括极限学习机（ELM），基因表达编程（GEP），自适应神经模糊推理系统（ANFIS）和广义数据处理方法（GMDH）来训练高精度的机器学习模型，使用详细的历史道路盐应用记录来处理一系列复杂的冬季风暴事件。这种新颖的模型可以根据道路天气信息系统（RWIS）提供的实时天气预报，在给定的冬季风暴事件中，更准确地规定在适当的时间间隔在道路和停车场上施用适当的道路盐量，以保护易受盐害的地区。我们将开发一个激光雷达盐分配测量反馈回路系统，使智能盐车能够自我校准，并将创建一个图形人机界面（HMI），通过记录和回放配置或通过实时控制门户网站设计，为操作员选择路线和启用gps的盐配置文件。最终目标是开发配备智能系统的下一代智能盐车，以优化盐应用计划，提高道路安全性，同时更好地保护已确定的盐脆弱区域，并以更低的成本。

More than 5-million tonnes of road salt is used for de-icing roadways and parking lots in Canada each year. These salts can have a significant impact on water quality. Both environmental and health & safety concerns are becoming progressively important in the provision of road salt application plans. This research project will develop novel models for prescribing the right amount(s) of road salt and at the right time(s) to be applied judiciously on roadways and parking lots for a given storm event. This model will be using both remotely-sensed and ground-based, real-time weather forecast data to prescribe the optimum road salt application rates and timing of the applications within the storm event to achieve road safety while minimizing the harm to the salt-vulnerable areas. We will use the Road Weather Information System (RWIS) to more accurately forecast the severity and type of the winter storm events. We will then use advanced Artificial Intelligence modeling techniques, including extreme learning machine (ELM), gene expression programming (GEP), adaptive neuro-fuzzy inference systems (ANFIS), and the Generalized Method of Data Handling (GMDH) to train highly accurate machine learning models, using the detailed historic road salt application records for a range of complex winter storm events. This novel model can help more accurately prescribe the right amount(s) of road salt to be applied at the right time interval(s) on roadways and parking lots for a given winter storm event based on the real-time weather forecasts, available through the Road Weather Information System (RWIS), to protect the salt vulnerable areas. We will develop a LIDAR salt dispense measuring feedback loop system to allow the smart salt truck to self-calibrate and will create a graphical Human Machine Interface (HMI) for the operators to select routes and GPS-enabled salting profiles via a record and playback configuration or via a real-time control web portal design. The ultimate goal is to develop the next generation of the smart salt trucks equipped with intelligent systems that would allow for optimized salt application plan for higher road safety while better protecting identified salt vulnerable areas and at a lower cost.