Abstract
This project aims to address the key remaining obstacles to general deployment of PedPal, a mobile app that assists pedestrians with disabilities in safely crossing signalized intersections. Tasks include developing sufficiently accurate localization to enable active monitoring of progress during crossing, generalizing the app to accommodate complex intersections, and developing a scalable V2I/P2I communication infrastructure.
Description
Overview
This project aims to further develop and deploy PedPal, a mobile app designed to assist pedestrians with disabilities in safely crossing signalized intersections. The PedPal app, which currently runs on an iPhone, was recently developed as part of the Federal Highway Administration’s (FHWA’s) Accessible Transportation Technology Research Initiative (ATTRI) program. [1] It utilizes "connected vehicle" technology to interact directly with signalized intersections, and through integration with the Surtrac real-time adaptive traffic control system, it gives pedestrians the ability to actively influence traffic control decisions to enhance their safety and mobility. Most basically, PedPal communicates a pedestrian’s crossing time requirements along with the desired crossing direction, ensuring that the pedestrian receives sufficient time to cross when the crossing signal is given while eliminating the need to locate and push a pedestrian call button at the corner to indicate presence. PedPal is also capable of dynamically extending the crossing time in circumstances where the user is moving slower than expected, and, if provided with pedestrian route information, PedPal can factor estimated pedestrian arrival times into the signal timing plans that are generated over time to streamline overall crossing time. As part of original PedPal development effort, structured user field tests of the technology were carried out periodically at selected intersections in the Pittsburgh Surtrac deployment by members of the local disability community, including blind individuals, wheelchair users, elderly individuals, and hearing-impaired persons. In these tests, the user response has been overwhelmingly positive, and the feedback obtained has served to refine and improve the app’s feature set and user interface. In addition, results have been produced that quantify the potential safety and mobility benefits over current intersection crossing practice. At the same time, our user testing involved some amount of “Wizard of Oz” experimental assumptions, and there are a few key technical issues that need to be resolved in order for the technology to be made available for general deployment.
In this project, we propose to undertake work to address these key remaining obstacles to deployment. First, we will develop localization techniques that enable real-time monitoring of crossing progress and detection of veering outside of the crosswalk, building on our recent success in using Bluetooth Beacons positioned at intersection corners in conjunction with smartphone localization capabilities to do automated corner identification. Second, we will generalize the current PedPal app, which is currently designed to operate at simple, 2-phase intersections (north-south and east-west), to handle complex intersections with multiple protected turning phases. Finally, with matching support from Rapid Flow Technologies Inc., the most likely PedPal deployment partner, we will develop a fully scalable, cloud-based V2I and P2I infrastructure for deploying the enhanced PedPal mobile app. Field test experiments will be conducted periodically to validate and refine each of these proposed capabilities. The technical scope of the proposed work is described in more detail below.
Technical Scope
Perhaps the biggest technical challenge to practical application of the PedPal mobile app in its current form is the lack of sufficient native localization accuracy in the iPhone (or for that matter any smartphone) to support corner identification, progress monitoring and veer detection, all of which are important functions for effective use. Most basic and most critical is corner detection. During the year 2 user test of the PedPal prototype, we implemented a Wizard of Oz user experience where the current intersection corner (NW, NE, SW, SE) was remotely communicated to the app whenever the user arrived at the intersection to initiate crossing or arrived at the destination corner at the end of a cross. (This information is required by the app to provide corner-specific orienting information to the user at the beginning of a cross, and to properly display crossing options once a user has completed a cross in one direction.) On post-test survey results, this user experience was identified as the single most important improvement over the Year 1 prototype, where the user was asked to tell the app whenever a crossing was started or completed.
Subsequent to this Year 2 user test but prior the end of the FHWA ATTRI project, an approach to corner identification based on introduction of additional infrastructure, specifically the placement of Bluetooth beacons on poles at each intersection corner, was investigated. An algorithm based on relative distance (range) readings obtained from various beacons was developed that provides a reliable approach to automated corner identification, and (realizing the Wizard of Oz generated behavior) allows the corner to be identified upon arrival eliminates the need for the user to input crossing starts and ends. This capability was incorporated into the final PedPal prototype produced by the FHWA ATTRI project.
Building on this approach, we will investigate the use of this additional Bluetooth sensing capability (1) as a means of enabling a basic progress monitoring capability, and (2) as a basis for detecting when a pedestrian is veering outside of the crosswalk. The former is essential to operationalizing PedPal’s ability to dynamically extend the crossing signal duration if slow progress is detected, and we believe that the same algorithmic approach taken for corner identification can be adapted to enable reliable progress monitoring by sensing source and destination signals. The latter capability of detecting when the user veers outside of the crosswalk may also be addressable with this same configuration of Bluetooth sensing. In this case, we will investigate triangularization of readings with an additional beacon not on the intersection crossing path. If this approach proves successful, further development of the PedPal app will be necessary to take advantage of it. Specifically, the MAP message specification will have to be extended to include crosswalk geometry information and the app extended to properly extract this information for use in detecting outside of the crosswalk veers. The PedPal user interface will also require extension to signal such events (e.g., by haptic vibration), and to properly indicate the necessary corrective direction.
A second prerequisite for general deployment of the PedPal technology is expansion of the types of signalized intersections that the app can handle. The current PedPal mobile app user interface makes a number of simplifying assumptions about the structure of the intersection, and currently only supports simple 2-phase intersections (i.e., consisting of just north-south and east-west traffic phases). A second focus of this research will be to generalize user interface capabilities to relax these assumptions and enable operation in complex intersections that have additional protected turning phases and/or a separate “all ped” crossing phase.
In addition to the above developments, we will work with Rapid Flow Technologies (RFT), the most likely PedPal deployment partner, to develop a scalable, cloud-based infrastructure for V2I/P2I communication. The FHWA ATTRI project resulted in development of two options for PedPal communication to the intersection: (1) via DSRC radios using an ARADA Locomate ME sleeve to interface with the smartphone, and (2) via cellular communication to the cloud (and then to the relevant intersection). As might be expected, user testing indicated a strong preference for use of a single hardware device, and since it is not currently economically feasible to provide a DSRC smartphone chip, a cellular, cloud-based option seems to be the best short-term option for deployment. Using the current FHWA ATTRI developed implementation of PedPal’s cellular communication option as guidance, RFT will provide matching support to develop a modular and fully scalable cloud-based V2I/P2I communication mechanism that can support broad deployment of V2I/V2P communication technologies.
References
[1] Smith, S.F., Rubinstein, Z.B, Marusek, J., Dias, B, and Radewick, H., “Connecting Pedestrians with Disabilities to Adaptive Signal Control for Safe Intersection Crossing and Enhanced Mobility: Final Report”, Technical Report FHWA-JPO-19-754, September, 2019.
Timeline
The above technical goals will be carried out according to the following quarterly timeline:
1st Quarter [July - September, 2020] – During the first three months of the project, work will focus on:
o Development of localization capabilities for reliable progress monitoring and veer detection, assuming use of Bluetooth sensors
2nd Quarter [October – December, 2020] – During the second quarter of the project, work will focus on:
o Generalization of the PedPal user interface to accommodate complex intersections
3rd Quarter [January-March, 2021] – During the third quarter of the project, work will focus on:
o Design and Implementation of scalable cloud-based V2I/P2I communication infrastructure (joint with RFT)
4th Quarter [April – June, 2021] - During the last quarter of the project, work will focus on:
o Testing of the new PedPal mobile app and the cloud-based communication infrastructure, and refinement of new app capabilities as appropriate
Strategic Description / RD&T
Deployment Plan
The most likely scenario for deployment of the PedPal technology is as an additional capability provided with the Surtrac traffic signal control system. Surtrac is currently marketed to municipalities by Rapid Flow Technologies, Inc., a CMU spinoff company (and a cost matching partner in the proposed endeavor). One baseline approach to deploying PedPal at this moment would be to offer it to municipalities as an extended feature that is provided free of charge as part of any purchased SURTRAC deployment. Municipalities could then offer the capability to their local disability community as a gesture of goodwill.
There are however additional considerations relating to how the PedPal technology is made available to prospective users. First, the fact that PedPal provides the ability to set the user’s speed and in return get the crossing time that is implied is a feature that is appropriate for pedestrians with disabilities, but not necessarily for any pedestrian that might have access to the app. Thus, it seems unrealistic that PedPal should be disseminated by simply making it available at the Apple App Store (or through Android’s equivalent site, once the PedPal technology has been ported). Instead it seems more reasonable to assume that access to PedPal will be mediated by some sort of user registration site, where eligibility can be determined in advance of access to download it. Second, it seems necessary to provide some sort of registry of PedPal enabled intersections in a given municipality or geographic area where PedPal has been made available, since an intersection that is not enabled will have no way of communicating that information to the mobile app. We will develop appropriate mechanisms for addressing both of these practical deployment issues.
Expected Outcomes/Impacts
We expect the proposed research plan to produce a deployable version of the PedPal mobile app for safe intersection crossing. To validate our technology development results, we will periodically carry out user testing at intersections in the Pittsburgh Surtrac deployment, following the same IRB-approved protocols that were used in the FHWA-ATTRI project. We will collect data to measure such factors as overall crossing time duration, number of missed cycles, and ease of use, particularly for those participants that have challenges operating a smart phone while crossing.
Expected Outputs
TRID
Individuals Involved
| Email |
Name |
Affiliation |
Role |
Position |
| vparimi@cs.cmu.edu |
Parimi, Viraj |
Robotics Institute |
Other |
Student - Masters |
| zbr@cs.cmu.edu |
Rubinstein, Zachary |
Robotics Institute |
Other |
Faculty - Research/Systems |
| sfs@cs.cmu.edu |
Smith, Stephen |
Robotics Institute |
PI |
Faculty - Tenured |
Budget
Amount of UTC Funds Awarded
$88419.00
Total Project Budget (from all funding sources)
$88419.00
Documents
| Type |
Name |
Uploaded |
| Data Management Plan |
data-management-plan.pdf |
June 29, 2020, 1:49 p.m. |
| Project Brief |
UTC-PedPal-slides.pptx |
Sept. 24, 2020, 1:34 p.m. |
| Publication |
Technology to Make Signalized Intersections Safer for Pedestrians with Disabilities |
Oct. 24, 2021, 8:51 p.m. |
| Final Report |
Final_Report_-_333.pdf |
Nov. 23, 2021, noon |
| Publication |
Integration of Automated Vehicle Sensing with Adaptive Signal Control for Enhanced Mobility |
April 5, 2022, 8:58 p.m. |
| Publication |
Technology to Make Signalized Intersections Safer for Pedestrians with Disabilities |
April 5, 2022, 9:01 p.m. |
| Publication |
Multiagent sensor fusion for connected & autonomous vehicles to enhance navigation safety |
April 5, 2022, 9:03 p.m. |
| Publication |
Paratransit Scheduling and Routing |
April 5, 2022, 9:04 p.m. |
| Publication |
Surtrac for the People: Upgrading the Surtrac Pittsburgh Deployment to Incorporate Pedestrian Friendly Extensions and Remote Monitoring Advances |
April 5, 2022, 9:05 p.m. |
| Publication |
Pedestrian Friendly Traffic Signal Control. |
April 5, 2022, 9:07 p.m. |
| Publication |
Incorporating Queue Dynamics into Schedule-Driven Traffic Control |
April 5, 2022, 9:08 p.m. |
Match Sources
No match sources!
Partners
| Name |
Type |
| Rapid Flow Technologies |
Deployment Partner Deployment Partner |
