The testbed includes 16 poles and each pole includes a wide array of sensors(including air quality, cameras, LIDAR, RADAR, and audio) and networking capabilities(SDR, DSRC, 802.11AC, and 802.11AD) shown in Table 1 and 2. Poles and industrial enclosures at each location allow for easy deployment of additional application specific hardware. 

All data generated within the testbed is ingested into a scalable, event-driven, publish/subscribe data integration platform. All data is accessible in real-time via APIs that are provided to partners. Prebuilt connectors can be utilized for continuous integration by popular data storage platforms such as HDFS, AWS S3, and Cassandra. These connectors run on our scalable connect cluster that can be deployed with no development, only configuration. Our localized data center allows low-latency, high-bandwidth connections that are necessary to meet the time critical application requirements. It also hosts operational data stores for analysis, reporting, and monitoring before the data lands in our data lake.



The physical infrastructure is supported by EPB’s gigabit fiber network. This will guarantee high throughput and low latency backhaul. In addition, we are working on several wireless technologies on the testbed, some are implemented and the rest are being investigated:

  1. DSRC: Access points are deployed throughout the testbed resulting in a single network for all devices to operate. Dedicated Short Range Communications (DSRC) is deployed throughout the testbed to allow high speed communications between vehicles and the roadside units and/or between vehicles with a range of up to 1 km.

  2. 4G, 5G, mmWave: We are exploiting other technologies to complement the DSRC technology to overcome the fact that DSRC’s throughput and delay performance degrades in large-scale connected autonomous vehicles. More specifically, we are investigating wireless technologies such as 4G Long-Term Evolution (LTE) and millimeter wave communication by using our testbed to evaluate and assess their performance. These technologies can be added to the testbed for a comprehensive assessment.

  3. UWB: Location-based applications highly depend on availability of GPS signal. Mapping applications for vehicle routing usually suffer in highly dense urban areas that GPS signal cannot penetrate due to shadowing. Ultra Wide Band (UWB) technology has been proposed for a precise positioning and navigation system. We are working on enhancing our testbed capability by adding UWB transmitters.

To ensure the security of the wireless communications, we have been working on implementing blockchain. Blockchain provides a secure and distributed database for scalable applications. The receiving unit can validate the authenticity of the received messages using blockchain database. Our testbed has processing units that will enable us to process and update the blocks for our distributed database. These edge units on our testbed makes the implementation of blockchain possible. We are in the process of investigating the practicality of using blockchain adaptation together with DSRC or other wireless communication (5G technologies) to ensure secure data transmission for future vehicle to vehicle or infrastructure communication. You can see a list of these devices with a simple description below:

Device Description
LimeSDR Handles I/O of numerous radio signal varieties
Locomate DSRC Dedicated short-range wireless communications
HackRF One Programmable peripheral for software defined radio like LimeSDR
Aruba AP 270 802.11AC Wireless Router with 2.4 and 5.0GHz frequencies
TPLink AD7200 802.11AD Wireless Router with 2.4, 5.0 and 60GHz frequencies
LoRa Gateway Physical Layer IoT Wireless Networking Component

Sensors & Processing

With the testbed's variety of sensors, it is possible to implement a variety of applications that easily utilize the data inputs that are obtainable. Many projects have been proposed and a few have been deployed using these sensors (such as the PurpleAir and YOLO Counter included on this site). These sensors and a basic description are listed below:

Device Description
Purple Air PA-II-SD Air sensor providing current air quality conditions
Axis P1448-LE 4K 30 FPS camera with variable FOV
Axis M2025-LE 1080P 30FPS camera with wide FOV
Sound Card + Preamp Mic Mic connected to Raspberry Pi for audio analysis and processing
RP LiDAR LiDAR sensor for bathymetric depth calculations and more
Banner Q240R RADAR Narrow-beam Fixed RADAR sensor

Edge Computing

To distribute processing and further decrease the inherent latency caused by distance, edge computing can be used. Edge computing will be used for audio processing using SoNYC, done via Raspberry Pi. An Nvidia Jetson TX2 will ideally be implemented for future projects requiring GPU utilization, such as object recognition or image processing. The testbed is equipped with these listed Edge Computing devices:

Device Description
Nvidia Jetson TX2 Edge Computing node for GPU-intensive computation or processing
Raspberry Pi (Multiple) Edge Computing node for simple processing or data transmission