Enabling Energy Demand Response with Vehicular Mesh Networks

Howard CheHao Chang, Haining Du, Joey Anda, Chen-Nee Chuah, Dipak Ghosal, H. Michael Zhang

Abstract

As utility companies transition toward a "Smart Grid," implementing real-time Demand Response (DR) becomes critical for balancing energy loads. However, collecting data from millions of disjoint household sensors presents a significant "last-mile" networking challenge. This paper proposes VMesh, a vehicular mesh network that utilizes city fleets—such as buses and postal vans—as mobile routers to bridge these gaps. By leveraging existing vehicle traffic, VMesh provides a low-cost, resilient communication backbone for dynamic energy pricing and monitoring.

Introduction

The implementation of Demand Response (DR) requires a reliable mechanism to transmit data from individual household sensors (monitoring gas, water, and electricity) back to a central utility office. Traditional sensor networks are often limited by low power and short transmission ranges, resulting in disjoint clusters. Relying on cellular networks or dedicated wired infrastructure for every household is prohibitively expensive. VMesh addresses this by turning vehicles already traversing urban streets into a self-healing, mobile transit network.

The VMesh Architecture

The VMesh system is composed of four primary components designed to facilitate seamless data flow from the edge to the central office:

  1. Sensing and Transceiving Units (STUs): Wireless sensors located in buildings that monitor utility consumption.
  2. Aggregation Points (APs): Static roadside nodes that serve as gateways between local STU clusters and passing vehicles.
  3. Mobile Routers (MRs): Vehicles equipped with wireless transceivers and storage that "carry" data across the network.
  4. Central Gateways (CGs): Fixed upload points, typically located at bus terminals or depots, where data is transferred to the utility's backbone.
VMesh system architecture and message flow within an urban environment.

To optimize data delivery, the system employs a "greedy" routing algorithm known as Most Forward Progress within Radius (MFR). To handle high sensor density, clustering techniques are used to mitigate MAC layer congestion. A key innovation is deflection routing: if a primary path is obstructed—for instance, by a bus stuck in traffic—the message is deflected to an alternative vehicle or path rather than being dropped.

Vehicle Diversity and Network Fidelity

A core insight of the VMesh proposal is that different vehicle types provide complementary coverage characteristics. While buses offer high-frequency, scheduled coverage along main arteries, postal vans or garbage trucks can provide "fine" coverage reaching every household, albeit at different periodicities.

Comparison of coverage, schedule, and cost characteristics across different vehicle types.

Vehicles act as "data mules" with virtually unlimited power and significant computing resources compared to static sensors. Furthermore, mobile routers offer a "stationary repair" advantage: maintenance is simplified because a malfunctioning router (the vehicle) eventually returns to a central depot for service.

Evaluation and Results

The authors evaluated VMesh using a custom-built simulation in a 40x40 grid environment. The study compared a NOTALK configuration (where mobile routers only communicate with fixed gateways) against a BROADCAST model (where routers can exchange data with each other).

Key findings include:

  • Data Success Rate (DSR): In the BROADCAST model, the DSR increased rapidly as transmission range grew, hitting a saturation point at a range of 6 units. In contrast, the NOTALK model showed a slower, linear increase.
  • Latency Reduction: In high-density scenarios, the BROADCAST method significantly reduced latency, with the delay dropping sharply and stabilizing once the transmission range reached 6.
  • Density as a Performance Driver: Mobile router density was found to be the dominant factor in system performance. The more vehicles participating in the mesh, the more resilient and faster the data collection became.

Conclusion and Future Work

While the current simulation uses a simplified "Random Walk" mobility model, it demonstrates the feasibility of using vehicular fleets for utility data collection. Future research will focus on more realistic mobility traces—such as scheduled bus routes—and testing the architecture in robust network simulators like Qualnet. Additionally, refining deflection routing to account for real-world urban traffic patterns remains a priority for making VMesh a viable solution for the future Smart Grid.

Summary updated

Abstract

Inter-vehicle communication is becoming increasingly important in recent years. Traditional research efforts on vehicular networks have been put into safety or infotainment applications. In this article, we propose a Vehicular Mesh network (VMesh) to inter-connect disjoint sensor networks and act as a data transit net work. Our VMesh networks are designed to provide a low cost, high fidelity, scalable, and fault resilient solution for data collections and disseminations. Preliminary simulation results based on a random mobility model show that the system has better performance when mobile routers are allowed to talk to peers. Mobile router density and transmission ranges dominate the simulation results. When the router density is high and/or transmission range is large, the data success rate will be high and latency will be small. There also exists a critical value of transmission ranges. There are no significant performance improvements once this critical value is reached.
BibTeX
@inproceedings{chang2004enabling,
  author = {Howard CheHao Chang and Haining Du and Joey Anda and Chen-Nee Chuah and Dipak Ghosal and H. Michael Zhang},
  title = {{Enabling Energy Demand Response with Vehicular Mesh Networks}},
  booktitle = {MWCN},
  year = {2004},
  doi = {10.1007/0-387-23150-1\_32},
}
Figure