Project: #14 Crowdsourced Traffic Calming Progress Report - Reporting Period Ending: March 31, 2018 Principal Investigator: Bob Iannucci Status: Completed Start Date: Jan. 1, 2017 End Date: Aug. 31, 2018 Research Type: Advanced Grant Type: Research Grant Program: MAP-21 TSET National (2013 - 2018) Grant Cycle: 2017 TSET UTC Progress Report (Last Updated: March 28, 2018, 5:25 p.m.) % Project Completed to Date: 85 % Grant Award Expended: 100 % Match Expended & Document: 100 USDOT Requirements Accomplishments The TrafficDot has been designed, prototyped and tested. We now have functional prototype devices ready for deployment in the CIty of Palo Alto. Each device consists of a hardened enclosure (milled polycarbonate) housing a circuit board with sensors (magnetometer, light sensor, humidity sensor) as well as three radio systems (LoRA, WiFi, and Bluetooth). We have successfully demonstrated extraction of clean magnetometer traces from the sensors and have built two prototype versions of signal classifiers. We've conducted structured experiments on a closed airfield (under the auspices of the US Naval Postgraduate School) with a variety of vehicle types and speeds, and have produced a labeled training- and test-set for researchers interested in using magnetometer data to classify cars. We have deployed a test LP-WAN network in the City of Palo Alto and we have conducted extensive field strength studies in and around the target deployment area. We have demonstrated that the network is adequate for the intended deployment and that the propagation and coverage criteria set out at the beginning of the experiment are within the capabilities of the network. We launched and are still conducting a study into the design of antennas for pavement-mounted sensors for traffic measurement. It is our belief that this work will be applicable across a variety of smart city sensing applications. We have built and tested a web user interface that allows us to both visualize traffic flows from our sensors and to perform basic "health checks" of the deployed sensor devices. We have developed a time synchronization protocol for LoRaWAN that is a simple addition to the current standard and existing open source libraries. We believe it can provide synchronization accuracy up to a few microseconds while maintaining our constraints on low power. We are currently submitting a paper on our work. Impacts The project has increased our appreciation of the challenges when it comes to outdoor sensing deployments. For an ambitious wide-scale project, there needs to be a balance between sensor hardware, robust packaging, reliable software and energy while overall keeping the cost economically low. In order to achieve this, it requires meticulous co-design of the said factors. Our study into LPWANs and wide-area deployments have uncovered several other challenges. While LPWANs enable low cost deployments and data collection, concerns such as management, maintenance and over-the-air updates rise under the constraints of limited communication rates. This involves heavier investments in the planning and design phases as current approaches assume high throughput and easy access to devices. Through our work in wide scale, outdoor sensing, we've crossed many other disciplines other than Traffic Calming. New relationships with USGS and Contra Costa County Transport Authority, we've begun considering fusing data across disciplines for previously disjoint efforts. Since the additional cost of a transducer on the hardware platform is low, we can imagine building devices that can sense many quantities other than those originally designed for. One such application is environmental sensing cross impact with traffic sensing. With environmental sensing and modeling, we open the possibility of predicting natural events that may impact traffic safety, such as floods and landslides, to enable early warnings. Conversely, we can monitor the impact of traffic and road management changes on the environment itself. On societal aspects, our partners in government have begun to think widely of using technology to improve the lives of their citizens. Our proposed work on sensing for traffic calming has sparked city officials for better ways to gather data and be more scientific in proving their suggested course of actions both to city council and affected communities. On a wider scale, our work has been influential in engaging cities to think about the Smart City and what other services they can provide to their citizens. Improved ability to quantify traffic flows (both spatial and temporal accuracies are improved by virtue of a dense sensor network) Other - A sensor board capable of detecting vehicles using a 3-axis magnetometer - Software libraries to program the sensor board and various onboard peripherals - A cloud-based LoRaWAN management platform for connecting gateways and devices to the network - A time series database for storage of sensor data and corresponding visualization dashboard - A labeled data set of 3-axis magnetometer traces for 8 vehicles traveling at 8 different speeds - A time synchronization module for LoRaWAN Outcomes New Partners Contra Costa County, Multitech, USGS, Pycom Issues Our initial sensor packaging had to be revised when the City of Palo Alto added new constraints after we had produced prototypes. We've begun redesigning the packaging to meet the new constraints while trying to maintain structural strength and low cost. This has proved challenging to do at prototype scale, with long cycle times and high cost per cycle. We are continuing to explore options for manufacturing that maintain a reasonable price point for the required constraints. Due to this, our deployments have been delayed to Q2 of 2018. Hackathon scheduling will follow at a date to be determined.