The key project objectives were: 1) the demonstration and evaluation of a low cost data acquisition system that would provide a rich data set accumulated from multiple vehicles, and 2) establish how this data, coupled with situational data from other sources, might be used to detect slippery road conditions, rough road conditions and potentially other beneficial uses by the Michigan DOT. This work is part of an on-going research initiative to identify how state DOTS will use and benefit from the large quantities of data to be generated by future connected vehicle programs and to assist in refining connected vehicle system requirements. This Slippery Road Detection (SRD) project is a focused effort on a specific set of uses but may nevertheless fit under the larger umbrella project called Data Use and Analysis Processing (DUAP) which is similarly identifying Michigan DOT functional requirements for each of the larger operational departments in MDOT. An analysis of the data requirements for the desired applications along with a survey of existing systems that could provide this type of data resulted in a data list of 31 parameters. Fourteen data items of this list can be sourced from the embedded OEM CAN network. Six can be obtained from a variety of personal devices and two specialty data parameters come from dedicated sensors (road surface temp and humidity). While CAN provide a rich vehicle performance set, many of the specific parameters needed by Slippery Road Detection (SRD) are considered proprietary by vehicle manufacturers and not available. Further, vehicle CAN message sets vary by OEM, model, and year, making each potential instantiation of the SRD system unique. This is why the Connected Vehicle (aka VII) program worked with SAE to standardize those specific messages needed for safety and mobility systems. Unfortunately, at this time, no manufacturer is supporting the standard J-2735 message set. Limited CAN information from current production systems that are available to this project have been negotiated with Ford Motor Company, therefore only Ford vehicles are currently compatible with the SRD system. Efforts are on-going in negotiating NDA agreements with other manufacturers through joint efforts of the national Safety Pilot program at UMTRI. Of the available systems surveyed, a Droid smartphone was adopted to meet the overall requirements where the balance between data set richness, cost and ease of deployment were critical. The availability of the internal 3-axis accelerometer provided the capability to monitor vehicle dynamics along with road roughness. The smartphone also provided the combined use of the device for vehicle monitoring when driving and business use by employees when not driving. This works well where smartphones are already provided to employees for their business use. The Droid smartphone uses a Bluetooth wireless link as the primary interface method between the device and an external sensor module and a CAN device. The Droid can work independently, however the data set is limited to on-board data sources (GPS, time, and accelerometer). Adding the CAN devices adds CAN signals and adding the Surface Patrol subsystem adds surface temp and humidity sensing. The easily installed system is now deployed on multiple MDOT vehicles. Over 13 G bytes of data have been accumulated over 30,000 miles. Most of the data was collected using the base system configuration. Only one vehicle is currently outfitted with the road surface temperature and humidity sensing, and two have the CAN interface. A number of Matlab visualization tools were developed to assist in evaluating vehicle performance, algorithm development and data reporting. The initial focus on vibration data coming from the Droid 3-axis accelerometer resulted in the development of a roughness measuring tool calibrated to measure PASER ratings. This has promise in assisting the MDOT in the continuous monitoring of road surface quality which is impossible to do using other methods. Other evaluations of vehicle performance data look at brake lockup, throttle position, vehicle speed, surface temperature and other parameters to assess whether a slippery condition can definitively be determined. Due to the limited amount of CAN data available to date, no conclusions can be drawn. The DUAP 2 program was launched in January 2012. DUAP 2 will start by interviewing MDOT organizations to determine their operational needs. From these needs, system requirements and specifications will be developed for data collection system designs. It is expected that the DataProbe system will an integral part of the final system and can be easily modified to accommodate additional data inputs that may be required. (Author/publisher)
Abstract