The California Connected Vehicle (CV) Testbed is located in the heart of the Silicon Valley in Palo Alto, California. The Testbed spans 16 consecutive intersections (from Medical Foundation Dr to Dinah's Ct) along a three-mile stretch of State Route 82 (El Camino Real) and is in the process of adding 15 more intersections (from Los Altos Ave to Grant Rd) that bring its length to 7 miles. It provides an operational environment where intersections and vehicles can communicate through wireless connectivity.

The Arterial Performance Measurement System is a tool that monitors the sensor network performance in the California CV Testbed. This system consists of three different analysis levels, which is shown below.

• Message/Data Element Level
As a first step, the system will check the data availability for the devices in the California CV Testbed. In particular, we focus on the real-time data availability for the detectors and intersection signals.
• Device Level
With stable data feeds from the field, the system further aggregates the raw detector and signal phase data into fixed time intervals or cycles. For the detector data, the system analyzes the quality of the data and help identify bad detectors in the Testbed. For the signal phase data, the system provides various statistics (e.g., phase duration, gap-out/max-out, and pedestrian activities) to assess the signal performance.
• Intersection Level
With the available detector and signal phase data, the system performs analysis to assess the traffic performance at the intersection level. Along El Camino Real, the system provides functions to visualize the coordination performance in different time periods (e.g., AM peak vs. PM peak). With the inputs of detector and signal phase data, the system conducts traffic state estimation and provides functions to visualize the formation and dissipation of traffic congestion at bottleneck intersections.

The architecture of the Arterial Performance Measurement System is shown in the figure below, which consists of the following key components:

• MySQL Database to store the detector and signal phase data.
• Mongo Database to store the arterial network and traffic estimation results.
• Data Aggregation component to aggregate the raw detector and signal phase data into fixed time intervals or cycles.
• Detector Health component to analyze the quality of the detector data.
• Data Filtering component to filter out the outliers in the detector data.
• Aimsun microsimulation model to host the static network and signal control information and for off-line traffic analysis.
• Network Builder component to provide TMDD input files for traffic state estimation.
• Arterial Traffic State Estimation component to generate estimates of traffic states at each intersection approach.
• Result Analysis and Visualization component to provide an interactive visualization tool (this website) for analysis.