We’ve all experienced the downfalls of public transportation: You wait 30 minutes for a bus during rush hour, then two arrive at once. You finally arrive at your stop, only to face a long walk home. Fixed-route buses can’t stop on every side street, and door-to-door on-demand buses are costly to run. What if one bus could run non-stop on the main road, then split into small units for the side streets? That is the idea behind modular transit – but how does it actually work?
A study led by Assistant Professor of Engineering Zhibin Chen presented at the International Symposium on Transportation and Traffic Theory (ISTTT), one of the most prestigious conferences in traffic flow theory, explores the viability of transportation solutions combining the flexibility of on-demand buses with the low operating cost of fixed-route buses.
Conventional fixed-route buses are relatively inexpensive to operate but might be inconvenient for some passengers whose origins or destinations are not close to bus stations. On-demand buses, by contrast, can adjust their routes to pick up/drop off passengers closer to their starting/ending points. This flexibility often comes at a higher operating cost for transit agencies, says Chen.
“Connecting a big station to small neighborhood stops is hard to coordinate in reality, but it’s much easier for modular buses,” he explained. “Multiple units can travel together over a long distance, then split up at the station and head in different directions.”
Chen and his team explored whether modular autonomous vehicles (MAVs) – made up of separate units that can connect and disconnect while in motion, allowing passengers to move between connected units – could offer a cost-effective and convenient solution.
The researchers designed an MAV-based transit system featuring a corridor service that runs through several zones with pick-up-and-delivery routes operating within each zone. Platoons of modular buses merge and split at hubs along the way, and passengers can board any MAV, transferring along the way to get to their destination.
The researchers built continuum approximation (CA) models with passengers spread evenly across the service area to calculate passenger cost (walking, waiting and in-vehicle travel time) and transit agency’s operating costs. They compared these to established fixed-route service and on-demand service models.

The study’s results suggest that MAVs can become particularly cost competitive on longer routes with high passenger demand. As demand increases, the cost per passenger of modular service can fall, making it more attractive than conventional bus services or on-demand transit.
The study also identified an interesting advantage for transit agencies with limited budgets. While conventional fixed-route buses may remain the only feasible low budget option, modular transit can offer a useful middle ground for transit agencies with a slightly higher budget, serving passenger demand while reducing walking, waiting, and travel-time costs. This flexibility comes from the system's ability to adapt vehicle capacity and routes to passenger needs, rather than requiring every vehicle to follow the same stopping pattern.

The study found several limitations facing the large-scale implementation of MAVs, including the requirement for reliable autonomous vehicles, relevant policies, and passenger safety needs. The research is based on a corridor model, whereas real cities have complex networks of intersecting roads and routes, adds Chen. He plans to further examine how vehicles could connect and separate across an entire urban network, as well as how passengers can move safely and efficiently between vehicles.
“I hope this research can help rethink how conventional bus services are designed,” Chen says. “If MAVs can eventually reduce operating costs while still meeting passengers’ needs, they could offer a new way to make the bus industry more sustainable.”
*banner image generated using ChatGPT


