California Partners for Advanced Transportation Technology (PATH) has been awarded nine projects by the Caltrans Research and Deployment Advisory Committee (RDAC) during this year’s proposal cycle, totaling $ 3,340,357.
“This is a high recognition of PATH’s contribution to meeting the California transportation goals to provide a safe, sustainable, integrated, equitable and efficient transportation system to enhance California’s economy and livability,” says Chief Operations Officer Zema Katsenlson.
Projects and PIs:
ADA Van as a Technology Demonstrator for Disabled Travelers
Alex Kurzhanskiy
Caltrans DRISI deserves to assume a leading position in transit automation, automated driving, connected automated vehicles (CAV), connected infrastructure and V2X. Presently, DRISI owns the connected infrastructure testbed at El Camino Real in Palo Alto, but does not have ideas how to effectively utilize it and demonstrate the benefits of the connected environment. This must change. Moreover, DRISI is supposed to showcase the emerging technologies and their applications for other Caltrans divisions and for policy makers. What will enable DRISI to master its role and fulfill its purpose?
Effects of Using Standardized Work Zone Data to Improve Work Zone Safety
Peggy Wang
Work zones play a key role in maintaining and upgrading our nation’s roadways. Unfortunately, work zone safety continues to be a concern for state transportation agencies across the US, for the safety of motorists who drive through the work zone and for workers who build, repair, and maintain our roadways. According to the national work zone safety data, in 2020, an estimated 102,000 work zone crashes happened, resulting in 44,000 injuries and 857 fatalities. In California, 88 fatal crashes happened, with a total of 96 fatalities. In specific, work zone intrusions are a growing concern for Caltrans due to their severe impact on the life of workers. The biggest share of work zone intrusion crashes occurs at lane closure operations.
Many transportation agencies face the challenge on how to gather and share work zone activity information with third parties effectively to improve work zone safety. Since 2019, USDOT has funded the Work Zone Data Exchange (WZDx) project (https://www.transportation.gov/av/data/wzdx(link is external)). This project focuses on the voluntary adoption of a basic work zone data specification, which enables Infrastructure Owner Operator (IOOs) to make harmonized work zone data available for third party use. Specifically, the project aims to get data on work zones into vehicles to help automated driving systems (ADSs) and human drivers navigate more safely and efficiently.
The PATH team conducted an automated vehicle (AV) Industry Survey Project in 2020, in which we surveyed and interviewed 20 companies regarding infrastructure needs for large-scale AV deployment. It was found that work zone and lane closure information is the most prominent digital information that AV companies need from the IOOs. Communicating the geographic information of work zones can greatly benefit the of Avs, thereby avoid work zone intrusions.
Prepare MMITSS for Deployment in California
Kun Zhou
The Multi-Modal Intelligent Traffic Signal System (MMITSS) has been developed under Connected Vehicle Pooled Fund Study (CV PFS) and several Caltrans sponsored research projects. The MMITSS software has been deploy in the California CV Test Bed1 along El Camino Real and is operational since 2018. To facilitate the MMITSS deployment in California, there is a need to document the existing MMITSS software that has been deployed in the California CV Test Bed, to improve MMITSS source code readability via better commenting, and to provide manual/user-guide for developers, management staff, and maintenance staff. There is also a need to develop a Software Over-The-Air (SOTA) update solution for updating the latest MMITSS software from Caltrans Traffic Management Center (TMC) to field intersections.
Evaluation of Artificial Intelligence-Based Video Data Collection Capabilities
Francois Dion
This project will evaluate the ability of emerging AI-powered sensors to satisfy Caltrans’ needs for general and freight traffic monitoring. This will be done by:
- Conducting a literature review on emerging AI-based detection technologies.
- Field-testing data collection from AI-based video detection sensors at select freeway and highway locations.
- Comparing data extracted from AI-based systems to data captured by nearby WIM, Census, or other legacy sensors.
- Evaluating the ability to extract gaps and other useful safety-related metrics.
- Evaluating the ability to distinguish various categories of trucks.
- Evaluating the ability to derive reliable origin-destination flow patterns or intersection turning movement counts from the collected data.
- Evaluating the ability to establishing heat maps for freight traffic movements based on OD data collected from the video cameras.
- Determining what legacy sensors could be replaced by AI-powered sensors and under which circumstances such a replacement would be beneficial.
- Developing recommendations on when and where to use AI-powered sensors.
Estimated timeline to complete the research: 18 months to 24 months.
Connected Intersections: Improving Intersection Safety and Mobility through High Resolution Video Detection and V2X Communications
Qijian Gan
This research aims to demonstrate a novel approach to connected intersections on the California CV Testbed that leverages both V2X communications and advanced video detection technologies and has great potential to simultaneously improve intersection safety and mobility and support future deployment of CAVs. In the California CV Testbed, AI-powered video and radar sensors were newly installed at four intersections for testing purposes. In this project, we will analyze the data from these advanced sensors, integrate it into testbed’s existing V2X communications and traffic control systems, and develop algorithms and applications to improve the intersection safety and mobility.
The proposed tasks in this research project are as follows:
- Quality analysis of the data from the AI-powered video and radar sensors.
- Development of a data management plan (e.g., standardized data formats, cloud storage schemes, data processing workflows, etc.) for future large-scale deployment. Algorithm development to detect traffic violations (e.g., red light running, jaywalking, and illegal turns) and potential hazards (e.g., near-misses and intersection blocking) at intersection approaches. Software enhancement to the existing V2X communications system to enable real-time broadcasting/multicasting information like presence of pedestrians and bicyclists on crosswalks, traffic violations, and potential hazards to nearby road users.
- Algorithm development to produce performance measures for multimodal traffic, including those from the ATSPMs and beyond (e.g., lane-based vehicle queues, pedestrian crossing time, pedestrian/bicyclist counts, lane blockages). Software enhancement to the existing Arterial Performance Measurement System (https://caconnectedvehicletestbed.org/apems(link is external)) to incorporate the new measures to support better evaluation of traffic signal performance.
- Field testing to validate the enhanced safety features in the testbed.
The estimated time to complete this research is 24 months.
SIDRA Calibration for Roundabouts in California
Alex Skabardonis
UC Berkeley PATH proposes an 18-month project, led by Professor Alexander Skabardonis, that will develop an automatic calibration for SIDRA roundabout model that will be used to study roundabouts in California. The purpose of this project is to study calibration of roundabout models in the software SIDRA Intersection for single lane, hybrid, and multilane environmental factors. The aim is to evaluate and compare different calibration methods, including creating an automatic calibration procedure based on optimization. For further use of the SIDRA Intersection model general settings would be useful, there for the second part of the purpose has been to test the possibility of general settings with a few models of roundabouts with different characteristics.
Three methods for calibration of the model will be tested in this project, one automatic based on optimization and two manual ones. Studies of the software, the model, and the theory will be basis for which parameters that are chosen for calibration. The first step is to find a suitable optimization algorithm to use for the automatic calibration. The choice of method depends on how easy the method is to use, in terms of number of control variables, and accuracy and robustness of results. The calibration parameters will become decision variables in the optimization problem. The estimated timeline to complete the research, 3 years.
Strategies for Improving Safety and Efficiency of Interactions Between Surface Traffic and Trunkline Transit
Joshua Meng
While traditional grade crossing can separate transit and surface traffic/pedestrians, its operation significantly impacts the surface traffic and likely will result in reduced capacity of BRT system due to the need to minimize the disturbance to the traffic.
This research plans to focus on:
- Understand that safety issues through analysis of hazardous activities involving interaction between transit and surface traffic,
- Study ways to apply advanced ITS technologies to detect and manage hazards while optimizing the operation of all modes.
- Compare the effects of treatments based on advanced ITS technologies vs traditional grade crossings vs. status quo.
This is an 18-month study with 6-month data collection at one or two intersections, and 12 months for data analysis and development of recommendations and report.
Use of Third-Party Data to Enhance Truck Monitoring in High-Traffic Networks
Francois Dion
This project will evaluate whether vehicle tracking data collected by third-party vendors could effectively be used to analyze truck movements in high-traffic urban/suburban networks. Using the Sacramento area as a test case, this will be done by:
- Assessing the representativity of truck movements captured by third-party vendors by comparing predicted volumes to measurements from fixed sensors.
- Determining the suitability for Caltrans’ needs of the truck classification(s) typically used by third-party vendors (usually based on gross vehicle weight).
- Developing a method for merging truck data provided by third-party vendors with data collected by Caltrans from fixed classification stations.
- Developing a method for assessing the reliability of extracted truck metrics.
- Developing a method to calibrate data from third-party analytical platforms
- Developing general guidelines on how to use third-party analytical platforms.
High Speed Data Distribution and Collection Prototype
Brian Peterson/Anthony Patire
This is a 12-month effort that will deliver a working small-scale prototype of California DTI. The primary concept of such a DTI is a system that spans from the sensor or infrastructure element on the road to a modern cloud based data and analytics system with tools that support real-time and long-term decision, policy, and results measurement and analysis.