Enabling App-Free On-Demand Transit Through Smart Bus Stop Infrastructure

A Hybrid Systems Approach to First-Mile Mobility in Low-Density and Regional Contexts

Mobility Is Not a Function of Density Alone. Mobility is too often framed as a byproduct of density, an assumption that prioritizes metropolitan areas and underserves small cities, regional hubs, and rural communities. Movement should not be afforded only by scale because it is a prerequisite for economic participation, social inclusion, and regional resilience. Cities and small towns must be connected in order to grow but also to function, so the ability to move reliably, affordably, and with dignity is of paramount importance. This white paper argues that modern transit systems need to decouple service quality from density by leveraging digital infrastructure, real-time systems, and demand-responsive operations that empower movement wherever people live.

Public transit agencies serving low-density and mid-sized communities face a structural mismatch between fixed-route transit (FRT) models and actual demand patterns. Peak-oriented schedules, long headways, and low vehicle occupancy result in high operating costs and poor rider experience, particularly during off-peak periods. 

Recent applied research conducted with Concordia University demonstrates that on-demand transit (ODT), when integrated directly into smart bus stop infrastructure, offers a scalable and inclusive solution to the first-mile problem.

BusPas introduces an on-demand interaction model that allows passengers to request service right at the bus stop without a smartphone, mobile application, or personal data account. This paper details the system architecture, dispatch logic, and empirical performance outcomes of this approach, drawing on Transportation Research Board (TRB) presentations and a peer-reviewed Transportation Research Record article. A particular focus is placed on relevance for regional cities where demand elasticity, accessibility, and cost control are critical policy objectives.

Structural Limitations of Fixed-Route Transit in Low-Density Environments

Fixed-route transit systems are optimized for predictable, high-volume corridors. In suburban, exurban, and regional contexts, these assumptions do not fit. Empirical studies show that low-frequency service leads to long wait times, poor reliability perception, and modal shift toward private vehicles, even when routes exist.

Agencies attempting to compensate by adding service often face diminishing returns, as marginal ridership gains do not offset increased operating costs. Moreover, fixed-route systems are inherently inflexible in responding to temporal demand variation. Off-peak periods (i.e. nights, weekends, and shoulder hours) are characterized by extremely low occupancy, yet vehicles must still operate to preserve network continuity. This results in inefficient capital utilization and elevated cost per passenger kilometre.

On-demand transit has emerged as a promising alternative, but most implementations are app-dependent, which introduces barriers related to digital access, privacy concerns, and friction related to user onboarding. The research presented here addresses these limitations directly.

System Architecture: On-Demand Transit Initiated at the Bus Stop

At the core of the BusPas model is the SmartSign display unit, an edge-computing device installed directly at existing bus stops. 

Each SmartSign unit integrates:

• An onboard edge processor
• User interaction hardware (physical button or touch interface)
• IoT sensors and connectivity modules
• Secure communication with cloud-based dispatch systems

Passengers can initiate a ride request by interacting with the SmartSign at the stop. This local, physical interaction eliminates the requirement for smartphones or applications which in turn significantly improves accessibility for seniors, low-income users, and privacy-conscious riders.

Data Flow and Cloud Integration

Once a request is triggered, the SmartSign transmits a structured data packet to the cloud. This data packet includes:

• Bus stop geolocation
• Timestamp of request
• Service zone identifier

The cloud layer functions as a central coordination platform, aggregating real-time data from SmartSign units and the on-demand vehicle fleet. Vehicles continuously report location, occupancy, and route state via APIs. The architecture enables real-time situational awareness across the system.

A conceptual representation of this interaction model is detailed in the ODT service framework diagrams presented in the TRB materials.

Dispatching Logic and Optimization

The Hybrid Transit Service (HTS) model integrates ODT with existing FRT by selectively replacing low-occupancy fixed-route buses with on-demand vehicles. Using simulation-based methods (SUMO), researchers identified buses operating below ridership thresholds and substituted them with dynamically dispatched vans distributed along the same corridors

Three scenarios were evaluated:

1. Current Situation (CS): Fixed-route only
2. Hybrid Route-Based (HRB): ODT deployed on select low-performing routes
3. Hybrid System-Based (HSB): ODT integrated system-wide

Reinforcement Learning–Driven Dispatch

A parallel study introduced ROUTE-Ride, a multi-agent rollout reinforcement learning algorithm designed to optimize vehicle-to-request matching under stochastic demand conditions.

Unlike static or greedy algorithms, ROUTE-Ride evaluates future state trajectories, balancing waiting time, detour penalties, and vehicle utilization.

The algorithm was validated using real Automatic Passenger Counting (APC) data from Laval, Canada, and implemented within the SUMO environment using TraCI integration. Results demonstrated significant computational efficiency and improved service quality across peak and off-peak periods.

Empirical Performance Outcomes

Across multiple simulations and datasets, the bus-stop-initiated on-demand model demonstrated consistent performance gains:

• Waiting Time:
  • ODT average waiting times under 2 minutes in hybrid scenarios, compared to over 8 minutes for FRT on comparable routes
  • ROUTE-Ride reduced waiting time by approximately 23% relative to baseline algorithms
• In-Vehicle and Total Travel Time:
  • Hybrid System-Based scenarios achieved a 53% reduction in in-vehicle time
  • Peer-reviewed results show 36% reduction in total travel time e 41% reduction in detour time compared to existing bus transit
• Fleet Efficiency:
  • Removing low-occupancy buses reduced capital and operating costs while maintaining or improving service levels
  • Increasing vehicle capacity beyond a threshold yielded diminishing returns, highlighting the importance of demand distribution rather than vehicle size alone

Crucially, these outcomes were achieved without door-to-door service, which typically increases detour time and operational complexity. The bus-stop-based pickup model strikes a balance between accessibility and system efficiency.

Policy Relevance for Regional Cities

Regional cities exemplify the conditions under which this model is most effective: dispersed origins, limited peak concentration, and strong reliance on transit for social and economic participation. 

The BusPas model enables municipalities to:

• Trigger service only when demand exists
• Maintain predictable rider experience through physical stops and real-time feedback
• Reduce underutilized fixed-route mileage
• Pilot innovation without overhauling fare or fleet systems

Because the system integrates with existing dispatch platforms and fare structures, it supports incremental deployment, which is critical for smaller agencies with constrained capital budgets.

Conclusion

The evidence presented through TRB research and peer-reviewed simulation studies demonstrates that on-demand transit does not need to be app-centric to be effective. By relocating the digital interface to the bus stop itself, transit agencies can deploy inclusive, adaptive, and cost-efficient mobility services that function across a wide range of urban and regional contexts.

Mobility should not be reserved for dense cores. With the right digital infrastructure, cities and small towns alike can be empowered by movement on their own terms, at human scale, and in real time.