An AI engine for operations
Real-time vehicle-to-trip matching, dynamic rebalancing, and demand forecasting — a system that keeps improving as operational data grows.
Capabilities
Three things we build for transit agencies.
A service simulator
Test zone boundaries, fleet sizes, and fare strategies in a full simulation of microtransit, buses, and traffic — before committing a single real vehicle.
A rider & operator platform
A friendly app for riders, a dashboard for planners and dispatchers, and an open data hub that publishes KPIs and anonymized datasets.
How it works
From "I need a ride" to a complete trip.
Open the rider app, set your pickup and destination, and request a trip — no phone calls, no guesswork.
The AI engine matches you to the right vehicle and shapes the route in real time — balancing your wait against everyone else's trip.
Vehicles are dynamically rebalanced so one is never far away. You track your ride live, and the driver gets a clear, optimized plan.
On-demand legs connect seamlessly to fixed-route buses and rail — booked as a single "Complete Trip" from door to destination.
Tap any step, or let it play.
How we work
Additive, not disruptive
Works alongside the systems agencies already run — never replacing them.
Explainable & auditable
Every recommendation comes with a plain-language reason; every decision is logged.
Locally controlled
The agency keeps final authority over every decision and full access to its data.
Open by default
The engines and data hub are open source; non-sensitive datasets are published openly.
A taste of our planning tool
Size a microtransit zone.
This is the kind of question our state-of-the-art planning tool answers for transit teams. Drag the controls to see how fleet size and demand change rider wait times and the share of trips served.
Turn on the AI engine
An illustrative public demo of our planning tool. Real deployments are sized with the full production model.
Trips served
{{ servicePct }}
Avg. wait
{{ wait }} min
{{ verdict }}
With {{ vehicles }} vehicles serving {{ demand }} requests a day, riders wait about {{ wait }} minutes and {{ servicePct }} of trips get a ride.
Who we are
A first-of-its-kind non-profit consortium.
Led by Vanderbilt University, PATH-TN brings together local university partners and transit agencies across Tennessee into a non-profit consortium — delivering a state-of-the-art planning-and-operations system, with every decision staying in local hands.
Dr. Abhishek Dubey
Lead PI · Vanderbilt University
Builds resilient, AI-driven systems for public transit and emergency response.
Research
Research behind the platform.
A sample of the research powering PATH-TN's planning and operations.
IEEE SMARTCOMP · 2026 · Best Paper
Dynamic Pickup-and-Delivery Routing with Early-Arrival Waiting Limits and Station Relocation
Microtransit routing that repositions vehicles and caps early arrivals to cut rider wait times.
Read the paper ↗ACM/IEEE ICCPS · 2026
Dynamic Vehicle Routing with Prompt Confirmation and Continual Optimization
On-demand routing that confirms rides instantly and keeps re-optimizing as new requests arrive.
Read the paper ↗Preprint · 2026
Column Generation for the Micro-Transit Zoning Problem
Designs efficient microtransit service zones — the question the demo above lets you explore.
Read the paper ↗ACM ICDCN · 2025
PDPTW-DB: MILP-Based Offline Route Planning for PDPTW with Driver Breaks
Offline route planning that builds mandatory driver breaks into the schedule so plans stay feasible.
Read the paper ↗IJCAI (Demo) · 2024
SmartTransit.AI: A Dynamic Paratransit and Microtransit Application
A live demo application that runs paratransit and microtransit together on one dynamic platform.
Read the paper ↗ACM/IEEE ICCPS · 2023
Mobility-On-Demand Transportation: A System for Microtransit and Paratransit Operations
A system for running microtransit and paratransit together — the operational core the platform builds on.
Read the paper ↗Funded By the Federal Highway Administration's ATTAIN program (Advanced Transportation Technology and Innovation), led by Vanderbilt University's Institute for Software Integrated Systems.
Good to know
Questions, answered.
A USDOT-funded, Vanderbilt-led partnership using AI to improve transit across cities in Tennessee. It builds planning tools, blends fixed-route and on-demand service, and designs smoother trips — while every operating decision stays with the local agency.
We run a realistic simulation of a city's transit — buses, microtransit, and traffic together — so agencies can try service designs like zones, fleet sizes, and fares in software before committing real resources.
Yes. Agencies keep full control of and access to their own data, and anything shared more broadly is de-identified and aggregated. The data hub publishes only anonymized datasets and KPIs.
The U.S. DOT Federal Highway Administration, through its ATTAIN program (Advanced Transportation Technology and Innovation), led by Vanderbilt University. See the grant on USDOT.gov.