Michigan Tech Research Institute (MTRI)

Unpaved Roads Assessment

Characterization of Unpaved Road Conditions Through the Use of Remote Sensing

An overview of the cooperative agreement between Michigan Technological University (MTU) and the US Department of Transportation Office of the Assistant Secretary for Research and Technology (USDOT/OST-R).

Effective Unpaved Road Condition Assessment and Management—The AURA System

The Federal Highway Administration (FHWA) reported in 2012 that there are over 1.4 million miles of unpaved roads in the United States, over one-third of the nation's road total (1). A 2000 study indicated that more than 512,000 miles of unpaved roads were on tribal lands, comprising 84 percent of total tribal roads (2). Unpaved roads, both gravel and unimproved roads, are a critical transportation resource that provide both rural and suburban transportation networks, including farm-to-table transition of food resources, getting children to school, and emergency routing after natural disasters, among many other uses. Unpaved road management is often the responsibility of local governments and transportation agencies, which are in need of rapid, repeatable methods that are cost-efficient and easily deployable in a budget-limited environment. In many cases, agencies do not have an up-to-date or complete inventory of the mileage of unpaved roads, making this an additional pressing need.

To address these needs, a project team led by the Principal Investigator Colin Brooks applied for and received a Phase III grant entitled “Characterization of Unpaved Road Conditions through the Use of Remote Sensing” from the USDOT Office of the Assistant Secretary for Research and Technology (OST-R) under the Commercial Remote Sensing and Spatial Information Program. This first phase of the project led to development of a cost-effective, accurate system for inventorying unpaved roads and assessing their condition, primarily using an Unmanned Aerial Vehicle (UAV) platform, now named the Aerial Unpaved Road Assessment (AURA) system. This included an on-board high-resolution 36-megapixel digital camera that combined 3D processing software, distress detection analysis algorithms, and GIS output that could be used for asset management in geospatial decision support tools.

For the second phase of the project, our team worked on a logical follow-on project to the previous effort so that our working prototype technologies can be taken from a useful, successfully demonstrated research-level tool to a commercially available, implemented system available to transportation agencies for objective unpaved road assessment on a day-to-day, as-needed basis. A particular focus was to make the system rapidly deployable for cost-effectively detecting deficiencies in the unpaved road components of the transportation system, including preventing of further damage to the transportation network through timely management of unpaved road assets.

A robust professional outreach program was planned to enable local, state, and tribal transportation agencies to understand the capabilities and values of the technologies for evaluating and managing their unpaved roads as part of their day-to-day methods. This was accomplished by continued work with a professional outreach firm, Integrated Global Dimensions, to reach more potential users; cooperation with one or more private-sector companies who expressed interest in helping with commercialization to make the system available on a day-to-day basis; continued close collaboration with transportation agencies such as partners in South Dakota; improvements to the distress detection software; and a series of field demonstrations in at least three states. Tasks to complete this goal enabled the AURA system to be a commercially ready product so that transportation agencies can more efficiently assess and manage this critical transportation resource.

  1. FHWA and USDOT. Highway statistics 2012. Highway Statistics Publications. http://www.fhwa.dot.gov/policyinformation/statistics/2012.
  2. Condition and performance of transportation serving federal and Indian lands. United States Federal Lands Highway Program. 2000. U.S. Dept. of Transportation, Federal Highway Administration, Federal Lands Highway Program. 18pgs.

 

A Look Back at the Project

For the January 2014 overview of the state of the project please view the summary of the Implementation Assessment of Unpaved Road Condition.

Task 1 Requirements Definition

August–October 2011

We will gather detailed information regarding the needs of the users and intended use of the system, features and ranges of values characteristic of unpaved road type and condition, concept and constraints of operation, performance, throughput, database requirements, and visualization of results and delivery needs. The expertise of team member Dr. Tim Colling, Director of the Michigan Local Technical Assistance Program and the Center for Technology and Training at Michigan Tech, was key in defining what characteristics have to be measured. MTRI staff member Lou Lambert, formerly the Bureau Chief of Planning for the Michigan Department of Transportation (MDOT) also helped provide input. We met with stakeholders, including our Technical Advisory Committee, to address real-world requirements of the transportation industry and reflect both their capabilities and limitations in funding or remote sensing expertise.

Outputs

  • A list of the system requirements, needed to proceed with the system analysis and design.
  • A broad but incisive characterization of the industry's current needs and expectations for this project, captured in the task deliverable.

Deliverables

Task 2 State of the Practice Assessment

November 2011

We reviewed prior work in unpaved road condition assessments, and included relevant methods in our trade study, including lessons learned from the Phase I USDOT/OST-R Commercial Remote Sensing and Spatial Information final report on "Monitoring the condition of unpaved roads with remote sensing and other technology" by Dr. Chunsun Zhang. We also tapped the experience of team members Brooks and Endsley on developing the Phase II State of the Practice content for the Bridge Condition Assessment using Remote Sensors study. Included will be a synthesis of the current processes of unpaved road condition assessment.

Outputs

  • A complete understanding of the current practices to guide the subsequent design phase, especially tasks 3 through 6 (Phenomenology, Sensor Selection, Platform Selection, and Algorithm & Software Development tasks).

Deliverables

Task 3 Phenomenology

November 2011–January 2012

For each useful feature characteristic of road condition, we tabulated the kinds of phenomena affected, the associated algorithms needed to analyze these phenomena, and ranked the general usability of each. Based on the team's experience, this included taking advantage of recent innovations in imaging technology that allow light's coherence and polarization to be used as assessment parameters as well. This helped enhance surface features that lend themselves to being detected by optical sensors, such as the largely surface-based features indicating unpaved road distress.

Outputs

  • An understanding of the remotely measurable phenomena associated with the set of road conditions needed for assessment and system design, including tasks 4 through 6.

Deliverables

Task 4 Sensor Selection

February–April 2012

We listed candidate commercial sensors likely to be able to meet the phenomenology needs. This included beginning the functional specifications of needed software, the size, weight, and power analysis, and the detailed specification of system components, including lenses, etc. As an example, if we have established requirements for the sensor standoff range, the field of view (FOV), the signal wavelengths and polarization, the resolution, and the frame rate, then these explicit requirements can be used to specify a sensor appropriate for unpaved road assessment. Standoff, FOV, and resolution can be used to compare sensors in terms of manufacturer specification. A determination of suitable commercially available sensors will need to be combined with filters (to get bands of interest and polarizations) and compared to possible collection rates. The options available to us were investigated and documented fully.

Outputs

  • A comprehensive, well documented list of the sensors and their characteristics that can be used to remotely sense the desired features, along with a rank-ordered list of those most likely to be useful in the final system.
  • A down-selected top-rated needed sensor(s) that will be used in the subsequent task, including documentation of justification for alternatives not pursued.

Deliverables

Task 5 Platform Selection

May–June 2012 and October 2014–January 2015

Given the sensor design, including its size, weight, and power requirements, and coupled with requirements on operation and performance, we selected an appropriate platform or platforms on which to mount the sensor system. The chosen platforms needed to be economical, easy to use with minimal training, and able to make the needed measurements as conveniently as possible. To keep the study feasible and timely, we proposed to evaluate a typical, manned, fixed-wing aircraft, as well considering possible UAV airborne platforms including fixed-wing, helicopter, and aerostatic (e.g. blimp) unmanned vehicles, to see if and when these platforms best meet the needs of the needs of the transportation user community, as evaluated through the Requirements Definition Task (#1) and the input of the Technical Advisory Committee. Either one, or both platforms, could have been selected through this process. We proposed that if a single platform solution is selected as the feasible option, a single practical version of that platform will be recommended by our team for development and deployment in Tasks 6 and 7.

If two platforms were selected, the chosen sensor modality from Task 4 would need to be modified to accommodate installation on both platforms. This will involve modest redesign to take into account the higher performance ratings of manned aircraft (altitude, speed, and maneuverability) vs. UAVs.
If more than one platform meets user needs, then a primary preferred platform and a secondary platform may be the resulting recommendation. We will perform a comparison of the any selected unmanned platform to the manned aircraft to understand when each system can each best be deployed to meet unpaved road assessment requirements. In order to maintain a focused budget and a two-year timeframe, we propose that if the recommendation is to use both the UAV and manned system, they will employ the same sensor modality, modified as needed for operation on both platforms. This will enable a timely and cost-efficient comparison of the relative strengths and weaknesses of these platforms, enabling USDOT/OST-R to be potentially prototype a multi-platform unpaved road assessment system that may be more capable than a single type of system.

Review and Update of Systems Requirements and Platforms – The URCAS will be reviewed to ensure that advances in platforms will be reflected in the commercialized product.

Outputs

  • An evaluation on which platforms, typical manned fixed-wing, a UAV, or both could meet the required sensing characteristics and be able to carry the previously specified sensor package. These platforms or platform will be used in the subsequent design and testing tasks.
  • A documented evaluation of the platform requirements, including criteria such as payload, time-on-station, expertise needed, mission support, and cost viability for a typical local or state transportation agency, as represented through the Technical Advisory Committee.
  • A comparison of any selected UAV platform sensor to a manned fixed-wing light aircraft version, in order to understand their relative strengths and weaknesses for unpaved road condition assessment, which will be continued as part of the Integration, Testing, and Demonstration Task (#7).
  • Review of platforms that could offer longer flight times over larger areas with increasing ease of use and deployment.

Deliverables

Task 6 Algorithm and Software Development

May–October 2012 and October 2014–July 2016

The software development comprises three major areas: image analysis and feature extraction/classification, database and decision support system software, and mission planning/control software, discussed below in Tasks 6.1 through 6.3.

Task 6.1 - Image Analysis

We tested and evaluated the image processing tools that enable extraction of relevant surface characteristics as defined in previous tasks. Both established internal tools and commercial software formed part of the evaluation. Algorithms included image morphology, segmentation, detection, and pattern analysis, and classification. The particular algorithms that will be most useful depended largely on the types, shapes, and sizes of the relevant surface characteristics defined in earlier tasks.

Outputs
  • Selection and recommendation of image processing tools needed to extract surface characteristics
  • Selection of algorithms to be applied through custom or existing image processing tools

Task 6.2 - Mission Planning and Control

We derived the unpaved road network as a major mission input using high resolution aerial images. Our primary anticipated input was the 30-cm resolution digital aerial imagery being provided as cost-share by the Southeastern Michigan Council of Governments. This base unpaved road network, which does not yet exist in detail, has been strongly indicated as a priority by SEMCOG, Michigan's Transportation Asset Management Council (TAMC), and local government agencies such as the Road Commission for Oakland County (RCOC), all cost-share partners on this project. These road type data, in the form a GIS layer, will form a major mission planning input that will allow route definition and flight path analysis. We will also develop or adapt mission planning software from existing commercial sources that will meet FAA guidelines for experimental aircraft, if needed based on platform selection. Remote sensing platform control software will assist the operator in navigation, flight profiles, and data collection. Tools will need to help an operator select an areas to measure, propose mission time and waypoints, needed mission supplies, and enable integration of the data collection and flight profile.

Outputs
  • Method for mapping the type and location of unpaved road locations for one or more priority study regions
  • GIS data layer showing unpaved road locations and types
  • Mission planning and control software capable of assisting the operator with end-to-end mission planning, maintaining safety margins, and optimizing mission performance; this will be incorporated into the integration task (#7).

Task 6.3 - Decision Support

We proposed to use the commercially-available RoadSoft GIS as Decision Support System (DSS) data and results sharing interface. Dr. Tim Colling led software customization and improvement efforts for RoadSoft and provided that role in this project, including how to effectively communicate alerts to users based on changes in road condition. We provided a working demonstration of how the unpaved road assessment and road type data generated through our methods can be shared through Roadsoft on a practical basis with the type of local, regional, and state agencies that need to make decisions using these types of data.

Outputs
  • A practical demonstration of RoadSoft GIS as the project's data sharing DSS
  • A user interface that will exploit a meaningful symbology to convey condition information to the end user
Deliverables

Task 6.4 Commercialization and Implementation

Adapting Remote Sensing Processing System (RSPS) into Commercial Readiness with GIS Extensions – Improvements to the 3D and distress algorithm components of URCAS, known as the RSPS, were made to add additional functionality to enable the work completed under the previous project to be commercially ready. This included the ability to analyze intersection geometry, drainage/flooding, haul roads and roadway safety audits. Also, focused additional development of the output XML road distress output data was completed, demonstrating how it can be integrated generically with GIS software such as ESRI ArcGIS and Michigan Tech’s commercially-available RoadSoft GIS Decision Support System (DSS) tool. By showing how the output data can be used in both existing road assessment-focused software such as RoadSoft, as well widely-available industry software such as ArcGIS, we enabled wider integration into asset management methods of more transportation agencies.

Outputs
  • A detailed list of the steps being taken to make the RSPS ready for commercial utilization.
  • Addition of additional detection features which include intersection geometry, drainage/flooding, haul road inspection and roadway safety audits.
  • Commercial-ready version of the URCAS Remote Sensing Processing System software, reviewed in the task deliverable.
  • Outreach for the use of geospatial XML distress data in GIS software to multiple transportation agencies.
  • A description of what steps were taken with GIS software to extend their capabilities to a nationwide audience when combined with URCAS.
Deliverables

Task 7 Integration, Testing and Demonstration

November 2012–September 2014

Integration of hardware and software allowed assessment of the operational characteristics of the unpaved road condition assessment system, the processing effectiveness, and the accuracy of the measurements. Included in this task is the development of field-test plans, as well as the final acceptance testing, detailing how the selected system(s) meet the program requirements. A final prototype will exhibit the most realistic concept of operations and practical utility of unpaved road assessment using the selected sensor and platform combination. Several miles of road, such as part of the 800 miles of unpaved gravel roads maintained by project partner Road Commission for Oakland County, will be characterized as part of this integrated system prototype. A performance evaluation will focus on the concept of operations, more detailed cost viability analysis, and practical utility of the selected system to assess unpaved road condition. Video Link: Field deployment and 3D reconstruction

Outputs
  • The outputs of this task are the fully-tested, operational prototype(s) of the system(s), with associated documentation.
Deliverables

Task 8 Reporting and Technical Advisory Committee

October 2011–September 2016

As a critical piece of outreach and project communication, we developed a project website to share results with the sponsor, TAC, and other stakeholders interested in improved efficient methods of assessing unpaved road condition. Included will be a clear link to the National Consortium on Remote Sensing in Transportation. The project website was maintained for the entire duration of the project.

We produced the needed quarterly reports, regular briefings to stakeholders on progress, and a detailed summary of the project as part of a final report. The final report will include accomplishments for each task, obstacles encountered, and how they were overcome. Key results and findings will also be described for each task. A narrative describing the design process from engineering requirements to the optimal solution will be included that delineates all of the options that were available for sensor and platform selection. Outreach and implementation efforts resulting from this project will also be described.

An additional key component of Task 8 is the formation of our Technical Advisory Committee, which will play a considerable role in defining the performance assessment methodology for this project, and will need to be approved by the Program Manager. We propose to have two in-person meetings, to be hosted at MTRI in Ann Arbor, MI. Periodic "WebEx"-style meetings can also be used to gain additional TAC input, on a schedule recommended by the sponsor.

Additional proposed outreach includes presenting to at least two major Transportation Research Board meetings, including working through proposer Brooks' leadership of the TRB Sensing Technologies Subcommittee. The Michigan Tech LTAP/CTT program also has much experience in effective outreach with state and local transportation agencies that will help with our outreach efforts, such as organizing the recent 2011 Michigan Bridge Conference.

Field Demonstration with State and Local Transportation Agencies - URCAS “in the field” demonstrations will be given to transportation agencies in multiple states; building from outreach efforts in Task 3 and interest gathered in states such as Nebraska, South Dakota, and Michigan. We anticipate repeating the successful format of the June, 2014 South Dakota field demonstration, with demonstrations of data collection procedures to local agency and private end users , technologies, assessment of distress detection algorithm results, and visualization of decision support integration. While demonstrations are budgeted for Michigan, Nebraska, and South Dakota, the project team will work with the project manager to select demonstration locations that enable the most effective, widespread outreach. Reporting and Technical Advisory Committee – Regular reports detailing the project’s current progress will be made to the sponsor and to stakeholders. In addition, deliverables and outreach materials will be posted to an updated project website. Technical Advisory Committee meetings will also be organized to gather input and share results based on the advice of the Program Manager. Updated Final Report – An updated final report will provide an in-depth review of the project’s methodology, findings, and accomplishments including outreach materials and details of field demonstration results.

Outputs

  • Project website, created and maintained for the entire duration of the project
  • At least two meetings with the Technical Advisory Committee, including reporting on TAC recommendations and feedback
  • Two or more WebEx-style interim meetings between annual TAC meetings to keep TAC up-to-date
  • Outreach efforts such as TRB presentations
  • Regular and clear reporting to the sponsor and stakeholder community
  • Exposure to multiple transportation agencies of the capabilities of URCAS for unpaved road asset management.
  • URCAS results collected and demonstrated for at least three planned states plus the location of the 2015 Local Roads Conference.
  • Summary of discussions with transportation agencies on how a deployed version of URCAS can help with their system deficiency detection and damage prevention needs.
  • Quarterly reports
  • Regular and clear reporting to the sponsor and stakeholder community
  • One or more meetings with the Technical Advisory Committee, based on recommendations of the Program Manager.
  • Updated project websites (research results focused and client-focused versions)
  • Draft final report for sponsor review
  • Final report that addresses sponsors comments

Deliverables

Task 9 Professional Outreach

October 2014–September 2016

Outreach events will be implemented to provide detailed, customer-focused descriptions of the URCAS system and how it can be integrated into unpaved road asset management procedures by state and local transportation agencies. Such events will include press releases, popular press articles, webinars, workshops, and professional outreach videos, building from recent efforts of V.Lefler that resulted in five new articles, a successful technology demonstration workshop, and a professional webinar. Depending on program manager direction, outreach to an international audience could be included; one trip for the PI to help present the work at the Asian Conference on Remote Sensing is included.

Outputs

  • Generation of promotional material such as flyers, demonstration videos, HTML email announcements and updates, implementation blog and webinars.
  • Generation of outreach documents such as a business intelligence report, creative brief, technical briefs, executive summary and SlideShare content.
  • Organizing field demonstrations for transportation agencies.

Deliverables

Task 10 Commercialization of URCAS for Day-to-Day Usage

October 2014–September 2016

Commercialization of URCAS for day-to-day use – Through partnering with the private sector, URCAS will be developed into a commercially-ready service provided to transportation agencies, in a time period that takes advantage of updated FAA regulations (expected in early 2016) for the UAV-based version that will coincide with this new project. The project team has already discussed commercialization efforts with Woolpert, Inc. of Dayton, OH, who would become part of the project team through this new funding for additional tasks. Woolpert would work with the project team to define the methods and potential market size of commercialization. The project team is open to working with other third-party service firms so that coverage of URCAS-type services could reach transportation agencies nationwide (see below under “Technical Barriers”).

Outputs

  • Generation of promotional material such as flyers, demonstration videos, HTML email announcements and updates, implementation blog and webinars.
  • Generation of outreach documents such as a business intelligence report, creative brief, technical briefs, executive summary and SlideShare content.
  • Organizing field demonstrations for transportation agencies.

Deliverables