Project “MIDAS: A Cyber Physical System for Proactive Traffic Management to Enhance Mobility and Sustainability “pertains to the exploration of an architecture for traffic management in a connected vehicles (CV) environment in future smart cities. The elements and infrastructures that the MIDAS architecture (see Figure 1) proposes are:

• Smart-phone based communication where CVs provide their current locations, traffic situations and, if available, destinations to a cloud of servers;
• A Cloud-based Cyber Infrastructure, which is secure, resilient, maintains privacy, that efficiently collects and manages data and concurrently performs image processing and traffic algorithms, while interfacing with the underlying communication system;
• PICT (position-image-communication with time stamp) devices, which are vital components of MIDAS, utilize cellular communications to monitor each CV , and guide CV drivers based on traffic-management algorithms that efficiently channelize vehicles using traffic controls, route advisories, and congestion pricing/incentives;
• Visual Computing, where the PICT devices use real-time lane and vehicle detection to identify the traffic situation within the roadway field of view, in order to place vehicles at the proper location within the traffic network;
• Online Traffic Estimation and Prediction Algorithms, which are vital for any proactive traffic management, provide the necessary prediction using novel models that combine mesoscopic and microscopic viewpoints of vehicles.
• Congestion Pricing Algorithms that, when predicted measures of performance such as heavy delays, emissions, etc., are not acceptable, recommend tolls/incentives and re-routing guidance to move towards more acceptable performances, where the PICT infrastructure provides the recommended congestion prices (or incentives) and parking prices to CVs and to the traffic management system;
• An Integrated MIDAS (Managing Interacting Demands and Supplies) Proactive Traffic Management System will utilize the interfaces with various developed algorithms, cloud/communication infrastructures, to provide proactive traffic management. It brings together data of CV’s movements via the PICT devices, and of the affected traffic systems, to provide proactive controls and advisories, to influence drivers to make better route choices, and to recommend congestion prices/incentives to effect travel demand.

The four primary investigators, Pitu Mirchandani (PI, ASU), Dijiang Huang (ASU), Baoxin Li (ASU), and Yafeng Yin (U. Florida), were able to (a) develop and test several of the components for MIDAS, (b) partially support three PhD dissertations (plus two expected in 2018), (c) publish a book in 2017 and a book chapter in 2015, (d) publish over 30 journal and conference papers, (e) obtain one patent, (f) collect 100 video clips available to researchers, and (g) develop two preliminary Android Apps. In addition the four investigators gave about 10 keynotes based on the project results, several invited university seminars and conference presentations.

INTELLECTUAL CONTRIBUTIONS

The research team has broken new ground in Operations Research (OR) and Traffic Engineering (TE) and Computer Science (CS), through the development of new models and algorithms, new cloud computing approaches and new visual computing methodologies. Most importantly, the team has developed a novel architecture for the MIDAS system that indicates a promising approach for proactive traffic management. Also innovative are the PICT devices that use a Cyber-Physical Infrastructure to connect vehicles and to manage drivers through a complex urban environment.

Specific pedagogical contributions to OR and TE areas are:

(A1) Online traffic estimation and prediction algorithms that use a new mesoscopic traffic model formulated in a “Lagrangian” coordinate system, with a recursive correction module that exploits the information provided by the PICT devices.

(A2) Congestion pricing algorithms use advanced theories in games and equilibria to develop recommendations on congestion prices (or incentives), including but not limited to (a) pricing on routes, (b) pricing on links, (c) pricing for subnetworks in the region, and (d) pricing for parking.

(A3) Proactive traffic management optimization algorithms that use real-time data obtained through the MIDAS/PICT infrastructure, and predictions from area A1, to recommend (a) proactive traffic signal controls, (b) travel route advisories, and (c) and congestion prices to influence travel demand so that the resulting traffic patterns, both recurrent and non-recurrent, enhance mobility and sustainability.

Specific pedagogical contributions to CS areas are:

(A4) A  cloud-based Cyber Infrastructure, which is secure, resilient while maintaining privacy, to perform image processing, execute developed algorithms and interface with the underlying communication systems.

(A5) Visual computing algorithms that use PICT data and images to identify, in real-time, CV’s lane position and detect the current vehicle movements and densities in camera’s field of view, which are subsequently used for proactive traffic management system.

BROADER IMPACTS

The results of this work will be useful to:

• State and local agencies that manage traffic to evaluate MIDAS’ potential impacts to congestion, environment and sustainability.
• Federal and regulatory organizations that need to evaluate safety of CVs and AVs operating on the roads.

Last but not least, the OR, TE and CS results from this project have permeated into the classroom where the investigators use cases and problems based on this work.

Last Modified: 05/30/2018
Modified by: Pitu Mirchandani