CHAPTER 1. INTRODUCTION

BACKGROUND

Traffic volume data are used by state departments of transportation (DOTs), metropolitan planning organizations (MPOs), and other local agencies in various analyses and functions related to transportation planning, design, traffic operations, safety, pavement design and maintenance, enforcement, environmental analysis, and construction, among others. State DOTs are required to report annual average daily traffic (AADT) every year to the Highway Performance Monitoring System (HPMS) for the full extent of mainlines, samples, and ramps on all federal aid facilities (Federal Highway Administration [FHWA], 2016a). In addition, the 2016 Highway Safety Improvement Program (HSIP) Final Rule requires states to have access to AADT for all paved public roads by 2026, including non–federal aid system (NFAS) roads (FHWA, 2016b). The NFAS includes rural minor collectors, rural local roads, and urban local roads. NFAS roads account for more than 75 percent of the total roadway mileage in the United States (FHWA, 2019). Counties and cities own around 88% of all NFAS roads (Tsapakis et al., 2020).

In addition to motorized traffic, there has been increasing interest over the last few years in monitoring non-motorized traffic. According to the 2022 Traffic Monitoring Guide (TMG), non-motorized traffic encompasses not just pedestrians and bicyclists but also travelers who use micromobility devices such as hoverboards, scooters, and e-bikes (FHWA, 2022). Like motorized traffic, the goal of monitoring micromobility traffic is to better understand traffic patterns, improve safety, and support decision-making.

AADT is typically calculated using data from continuous count stations (CCSs) that are permanently installed at select locations and operate 24 hours a day. Due to the high purchase, installation, operation, and maintenance costs of CCSs, agencies often use portable traffic recorders (PTRs) to conduct short-term or seasonal counts at segments where CCSs do not exist. The goal is to accurately expand the short-term counts to AADT estimates using appropriate adjustment factors.

Considering the (a) federal requirements mentioned above, (b) high data collection costs associated with traditional traffic monitoring equipment, and (c) need to collect and use traffic volume data in various applications and functions, transportation agencies try to obtain traffic counts from other traffic detection assets such as loop detectors, magnetometers, radar sensors, pan-tilt-zoom (PTZ) cameras, and traffic detection cameras. These assets are often purchased for purposes other than traffic volume estimation. One or more detection systems may be installed together to generate data that can be used not only for Automated Traffic Signal Performance Measure (ATSPM) estimation and signal control optimization (An et al., 2018) but also for traffic volume estimation, including TMCs (Li et al., 2019; Li and Wu, 2021).

Over the last few years, several agencies have started to extract volume data from traffic signal assets. The data include motorized volumes, non-motorized volumes, and TMCs. They are obtained from different types of signal control assets manufactured by various companies. Some assets can count one traffic mode (e.g., only motorized traffic), while others can count multiple road user types. Under certain circumstances, these signal assets can be used to extract volume data to calculate AADT and develop seasonal adjustment factors and other traffic parameters.

SCOPE AND ORGANIZATION

The scope of this guide is to provide transportation agencies with information about extracting motorized and non-motorized traffic volume data from existing signal assets. For completeness, the guide covers both commonly used signal assets and less commonly used but promising types of signal equipment. The guide is organized as follows:

• Chapter 2—Inductive Loop Detectors.
• Chapter 3—Video-Based Systems.
• Chapter 4—Microwave Radar Sensors.
• Chapter 5—Light Detection and Ranging (LiDAR) Sensors.
• Chapter 6—Infrared Sensors.
• Chapter 7—Magnetic Sensors.
• Chapter 8—Ultrasonic Sensors.
• Chapter 9—Other Technologies.
• Chapter 10—Data Management.

For each type of equipment, the guide offers information about its operation, strengths, weaknesses, count accuracy, cost, and recommended practices for improving the accuracy of traffic volume data. Each chapter provides information on how the respective technology applies to motorized and non-motorized traffic so that users focused on a specific mode can easily find the relevant details. For clarity, the strengths and weaknesses of each technology are presented for both traffic modes combined, as well as separately for motorized and non-motorized traffic where distinctions exist. Additionally, recommendations specific to non-motorized traffic are provided separately to highlight key considerations for accurate data collection. The guide also provides recommendations for obtaining, storing, transmitting, and managing count data for traffic monitoring purposes in Chapter 10.

BENEFITS

The ability to continuously extract counts from traffic signal equipment that already exists at hundreds (and in many cases thousands) of locations within a state transportation network creates a value proposition, which agencies have started to explore. Many signalized intersections are on lower roadway functional classes that may not be continuously and extensively monitored by CCSs. Extracting counts from traffic signal equipment offers several benefits for both traffic operations and traffic monitoring, as described below.

Traffic Operations

• Real-time traffic information: Traffic signal equipment can provide real-time traffic volume data and other information to agencies, thus enabling quicker response to traffic conditions and signal optimization, and in turn improving travel times and reducing congestion.
• Improved signal timing: Accurate volume data from traffic signals can be used to refine and adjust signal timing plans to accommodate real-time traffic demands, improving traffic flow.
• Integration with other data sources: Traffic data from signal equipment can be integrated with other data sources, such as global positioning system (GPS) tracking and advanced traffic management systems (ATMSs), where volume data can be integrated into broader traffic management strategies to optimize traffic conditions across a network.

• Intersection performance measures: Volume data from signals can help engineers calculate key performance measures (e.g., level of service) to assess how well intersections are performing and identify areas for improvement such as congestion points or underutilized lanes.
• Reduced manual data collection: Leveraging volumes from traffic signals can significantly reduce the need for conducting manual traffic counts, saving time and resources.
• Data-driven signal maintenance: By monitoring traffic volumes, signal engineers can more effectively plan and prioritize maintenance activities, ensuring optimal signal performance where it is needed most.
• Improved safety: Traffic volume data can be used in safety analyses to identify hazardous intersections and inform safety improvement measures.

Traffic Monitoring

• Cost reduction: Using signal assets to collect volume data can reduce the number of short-term counts needed in the field, as well as capital and operational costs related to the purchase, installation, and maintenance of PTRs.
• Geographical data coverage expansion: Leveraging the widespread deployment of traffic signal infrastructure, especially on lower functional class roads with limited or no other monitoring equipment, allows the collection of volume data at numerous locations without the need for separate traffic counting devices.
• Continuous count program expansion: Traffic signal equipment can continuously collect traffic volumes, which can be transmitted, stored, and used by existing continuous count traffic monitoring programs, allowing the latter to expand without the need for purchasing and installing new permanent traffic recorders.
• Traffic statistic calculation: Like CCSs, traffic signal equipment can collect and transmit data over extended periods, enabling the analysis of historical traffic patterns and the calculation of traffic statistics such as AADT, seasonal adjustment factors, directional factors, peak-hour factors, etc.
• Cost-effective monitoring: Utilizing existing traffic signal infrastructure for data collection can be more cost-effective than deploying and maintaining new monitoring equipment.
• Improved safety of field personnel: Reducing the number of short-term counts minimizes field personnel’s exposure to various roadway safety risks when installing portable recorders, thereby enhancing their safety.
• Intersection-level accuracy: Volume data from signalized intersections can offer a detailed, intersection-specific view, which is important for understanding vehicle flows and turning movements at critical nodes in the network.
• High temporal data resolution: Signal equipment often collects data at short intervals (e.g., every 5 or 15 minutes), which can be used for granular analysis of traffic trends (e.g., hourly, peak-period analysis, etc.).

TARGET AUDIENCE

This guide is intended for state and federal agencies, MPOs, local governments, and other transportation entities that own, operate, or are interested in traffic signals and the data collected from signal equipment. The primary audience includes traffic operations and monitoring staff, such as managers, traffic engineers, data analysts, database administrators, planners, data technicians, and IT personnel involved in the installation, operation, data collection, processing,

analysis, development, integration, archiving, publication, or reporting of traffic statistics like AADT. A secondary, broader audience includes those who may not work directly in traffic operations and monitoring but use traffic volume data for other transportation-related functions (e.g., safety, environmental analysis, pavement design and management, urban planning, public transit planning, and emergency services) or non-transportation purposes (e.g., real estate, billboards, insurance, tourism, logistics and supply chain management, economic and business development, event planning, utility management, etc.).

RESEARCH PROJECT

This guide is the result of National Cooperative Highway Research Project (NCHRP) Project 03-144: Leveraging Existing Traffic Signal Assets to Obtain Quality Traffic Counts and Enhance Transportation Monitoring Programs. The objectives of this research were to (a) determine the feasibility of using existing or enhanced traffic signal equipment to collect, store, and disseminate data for purposes other than traffic operations, particularly for traffic monitoring programs; (b) determine the suitability of traffic count data from already installed and existing traffic assets for this purpose; and (c) develop effective practices for obtaining and integrating traffic counts from existing traffic assets.

As part of the research, traffic volume data from various types of existing signal equipment were validated. The data were extracted from eight types of signal equipment: inductive loop detectors, video-based systems, radar sensors, LiDAR sensors, thermal sensors, vehicle-to-everything (V2X) sensors, magnetometers, and push buttons. These devices were installed at 376 intersections across 17 states and one Canadian province, covering over 1,250 intersection approaches and 2,600 lanes. Key findings and evaluation results from the NCHRP 03-144 project are included in this guide, while the project’s final report offers more detailed information on all research activities and outcomes. The researchers used Grammarly to improve the writing quality during the revision process of the guide and the final report.

TECHNOLOGY COMPARISON

A high-level comparison of the different signal detection and counting technologies discussed in this guide is provided in Table 1. The modes columns indicate whether technology is typically able to count motorized traffic, bicycles, and pedestrians. The next column indicates the number of sensors that are needed, with low meaning that one to two sensors can cover a typical intersection, medium meaning one sensor per approach, and high meaning one sensor per lane. The cost column is based on the average cost per intersection reported by survey respondents in NCHRP 03-144. The accuracy columns are based on the validation results of NCHRP Project 03-144. The qualitative rankings reflect the accuracy of the best intersection using existing dual-purpose sensors (i.e., used for both signal detection and data collection).

The accuracy of non-motorized traffic counting is low, because signal equipment validated in this project was originally installed by transportation agencies solely for vehicle detection and signal control purposes, not for collecting non-motorized traffic volumes. However, the accuracy of non-motorized counts can be significantly improved if the equipment is properly installed, calibrated, and validated to both operate traffic signals and collect non-motorized traffic counts. Since existing signal equipment was not initially designed or optimized for counting non-motorized traffic, its validation in NCHRP 03-144 was essential to identify potential sources of error and develop strategies to mitigate them. Many of these strategies are presented in the following chapters of this guide.

Table 1. Signal Technology Comparison.

Technology	Modes		Number of Sensors	Cost	Accuracy
	M	NM			M	NM
Inductive loop detectors			High	$$	Good	Poor
Video-based systems			Low/Medium	$$	Good	Poor
Microwave radio sensors		N/A	Medium	$	Good	N/A
LiDAR sensors			Low	$$$	Good	Poor
Infrared sensors			Medium	$	Fair/Good	N/A
Magnetic sensors		N/A	High	$$	Good	N/A
Ultrasonic sensors		N/A	Medium	N/A	N/A	N/A
Hybrid sensors			Medium	N/A	Good	N/A
Pedestrian push buttons	N/A		High	$	N/A	Poor

N/A = not applicable.

Number of sensors: Low < 4, Medium = 4, High > 4 per intersection.

Cost: Low ($) < $30k, Medium ($$) $30k - $40k, High ($$$) > $40k.

Accuracy: Poor: WMAPE > 20%, Fair: 10<WMAPE≤20%, Good: WMAPE ≤ 10%. This rating reflects the accuracy of existing signal equipment validated in NCHRP Project 03-144. The accuracy of non-motorized traffic counts is low because most of the equipment was originally installed by transportation agencies for signal operation and motorized traffic detection, not for counting non-motorized traffic.