⚡ electrical engineering

A MEC-Based Optimization Framework for Dynamic Inductive Charging

This paper proposes a Model Predictive Control framework leveraging edge computing to optimize dynamic inductive charging power allocation in urban traffic, demonstrating through Istanbul-based simulations that prioritizing vehicles with critical battery states significantly improves both resource utilization and user satisfaction compared to uncoordinated distribution.

Original authors: Emre Akıskalıoğlu, Mustafa Atmaca, Lorenzo Ghiro, Giovanni Perin, Renato Lo Cigno

Published 2026-05-01

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Emre Akıskalıoğlu, Mustafa Atmaca, Lorenzo Ghiro, Giovanni Perin, Renato Lo Cigno

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a busy highway where electric cars don't just drive; they also "drink" electricity wirelessly from the road itself as they move. This is called Dynamic Inductive Charging (DIC). Think of the road as a giant, invisible water hose, and the cars as cups trying to fill up while they race down the street.

However, there's a catch: The "water hose" (the power grid) has a limited flow. If too many cars try to drink at once, the water pressure drops, and some cars might not get enough to finish their trip. This is the problem the authors set out to solve.

Here is the breakdown of their solution, using simple analogies:

The Problem: The "Free-For-All" Buffet

Currently, if you drive on this special road, you just grab power whenever you can.

The Issue: If a car with a nearly empty battery (a "thirsty" car) and a car with a full battery (a "full" car) are side-by-side, they both grab power equally.
The Result: The thirsty car might still run out of juice before it reaches its destination, while the full car wastes power it doesn't need. It's like a buffet where everyone grabs a plate at the same time, leaving the hungriest person with nothing.

The Solution: The "Smart Traffic Cop" (MEC & MPC)

The authors propose a new system using Mobile Edge Computing (MEC) and Model Predictive Control (MPC).

The Analogy: Imagine a super-smart traffic cop standing on a hill (the "Edge") who can see the whole road ahead. This cop talks to every car via a walkie-talkie (V2X communication).
How it works:
1. Asking for Needs: Every car tells the cop, "I have 10% battery left and need to get to 80%," or "I'm already at 90%, I just need a top-up."
2. Predicting the Future: The cop doesn't just look at who is there right now; they look at where the cars are going to be in the next few seconds (the "Predictive" part).
3. The Decision: The cop acts like a fair referee. If the power is scarce (like a drought), the cop says, "Okay, the car with the empty tank gets the biggest hose right now. The car that is already full gets a tiny trickle or waits a moment."

The Experiment: A Real-World Test

The researchers tested this idea on a 10-kilometer stretch of road in Istanbul, Turkey. They used a computer simulation (a digital twin of the real city) to create three different traffic scenarios:

Night (Low Traffic): Few cars. Everyone gets enough power easily. The new system doesn't change much here.
Lunch (Medium Traffic): Cars are coming and going. The old system starts to struggle, but the new "Smart Cop" keeps things balanced.
Rush Hour (High Traffic): The road is packed. This is where the magic happens.

The Results: Fairness Over Speed

When the road was packed and power was running low, the difference was clear:

The Old Way (No Cop): Power was split evenly. Everyone got a little bit, but the cars with empty batteries were left dangerously low, risking them getting stranded.
The New Way (Smart Cop): The system prioritized the "thirsty" cars. It took power away from the cars that were already full and gave it to the ones in trouble.
- The Outcome: Fewer cars were left stranded with empty batteries. The "thirsty" cars were much happier, even if the "full" cars got slightly less than they asked for.

Why This Matters

The paper argues that for wireless charging roads to work, we can't just build the hardware; we need the software brain to manage it. Without this smart management, the system is inefficient and unfair. With it, we can ensure that the cars that really need power get it first, making electric vehicles more reliable and reducing "range anxiety" (the fear of running out of power).

In short: They built a digital "traffic cop" that ensures the hungriest cars get fed first when the food supply is low, preventing anyone from getting stranded on the road.

1. Problem Statement

The widespread adoption of Battery Electric Vehicles (BEVs) is hindered by "range anxiety" and long recharging times. Dynamic Inductive Charging (DIC) offers a solution by enabling wireless power transfer while vehicles are in motion, potentially reducing battery size and costs. However, DIC infrastructure faces two critical challenges:

High Cost & Power Constraints: DIC systems are expensive to deploy and operate under strict power budgets.
Uncoordinated Resource Allocation: Current literature focuses on single-vehicle performance or infrastructure planning. There is a lack of system-level strategies for allocating limited power among multiple vehicles with heterogeneous needs.
The Core Issue: Without intelligent coordination, uncoordinated charging leads to suboptimal resource utilization (when demand is high but power is available) and critical user dissatisfaction (when demand exceeds capacity, leaving vehicles with depleted batteries unable to finish their trips).

2. Methodology

The authors propose a Model Predictive Control (MPC) framework integrated with Mobile Edge Computing (MEC) and Vehicle-to-Everything (V2X) communications to optimize power allocation dynamically.

A. System Architecture

Scenario: A realistic 10 km urban road segment in Istanbul, Turkey, modeled in the SUMO traffic simulator. The road features 10 DIC "stripes" (segments with embedded coils) across two lanes.
Communication: Vehicles transmit battery status (State of Charge - SoC), capacity, desired exit SoC, and route data to an ITS-PM (Intelligent Transportation System Power Management) unit hosted on a MEC server. The server calculates optimal power assignments and sends them back to vehicles.
Control Loop: The system operates in discrete time steps ( $\Delta T = 5$ seconds) over a prediction horizon ( $K$ steps).

B. Optimization Formulation

The core of the methodology is a Convex Quadratic Program (QP) solved at each time step to minimize a cost function ( $J_t$ ) subject to physical and operational constraints.

Decision Variables:
- $P_{v,k}$ : Power allocated to vehicle $v$ .
- $P_{s,k}$ : Power assigned to stripe $s$ .
- $b_{v,k}$ : Predicted battery SoC.
Constraints:
- Battery Dynamics: SoC evolution based on consumption and charging efficiency.
- Power Bounds: Limits based on coil capacity, vehicle pad limits, and total system budget ( $P_{tot}$ ).
- Stripe Constraints: Power must be distributed among stripes, allowing for dynamic rebalancing (up to a factor $\rho$ ) between adjacent stripes.
Objective Function ( $J_t$ ):
The function minimizes three components:
1. Battery Tracking Error: Minimizes the squared difference between predicted SoC and the desired target SoC ( $b^{des}_v$ ).
2. Urgency Weighting: Prioritizes vehicles with a large gap between current and desired SoC and those with short remaining travel times ( $\tau_v$ ).
3. Revenue & Efficiency: Includes terms for operator revenue (energy price) and minimizing the gap between assigned stripe power and actual vehicle absorption to maximize utilization.

3. Key Contributions

MPC-Based Power Allocation Algorithm: A novel algorithm that dynamically balances individual vehicle urgency, predicted trajectories, and system-wide capacity constraints to maximize fairness and efficiency.
Communication Protocol Design: A defined protocol for exchanging minimal but critical data (battery levels, routes, power assignments) between vehicles and the infrastructure using existing V2X standards (ETSI ITS-G5) augmented with specific power management messages.
Open-Source Simulation Framework: The authors released a SUMO-based simulation tool that models realistic DIC constraints (stripe topology, power budgets, traffic variability) to enable reproducible research in this domain.

4. Results

The framework was evaluated against a Benchmark (BENCH) strategy (static, uncoordinated allocation) under three traffic intensities: Low ( $\lambda=5$ ), Medium ( $\lambda=12$ ), and High ( $\lambda=20$ vehicles/min).

Low Traffic (Abundant Resources): Both BENCH and OPT strategies performed similarly, satisfying all vehicle requests. Coordination offered marginal benefits here.
Medium Traffic (Oscillating Demand):
- BENCH: Struggled during peak demand, failing to fully utilize available power capacity due to static allocation.
- OPT: Achieved near-perfect resource utilization by dynamically rebalancing power between stripes based on real-time demand.
High Traffic (Resource Scarcity):
- Resource Utilization: Both strategies saturated the total power budget.
- User Satisfaction ( $\phi_v$ ): This was the critical differentiator.
  - BENCH: Resulted in a "heavy tail" of critically unsatisfied vehicles. Uniform distribution meant many vehicles left with insufficient charge, risking range anxiety or emergency stops.
  - OPT: Implemented a fairness-first policy. It aggressively prioritized vehicles with depleted batteries (Vehicles Under Test - VUTs) at the expense of well-charged vehicles.
  - Outcome: The Cumulative Distribution Function (CDF) showed that OPT significantly reduced the number of critically unsatisfied vehicles (improving the lower tail of the distribution), ensuring that the most urgent needs were met even if it meant some less-critical vehicles received slightly less than their maximum possible charge.

5. Significance and Future Work

System-Level Viability: This is the first work to analyze DIC systems at the traffic flow level, proving that intelligent power management is essential for economic and operational viability.
Economic Impact: By maximizing user satisfaction during peak loads and enabling dynamic pricing, the framework supports the economic sustainability of DIC infrastructure.
Battery Reduction Potential: Efficient DIC could allow for smaller, cheaper vehicle batteries, further accelerating BEV adoption.
Future Directions: The authors suggest future research into comparing strategies across different environments (highway vs. urban), optimizing communication protocols for 5G/6G, and investigating the impact on battery manufacturing requirements.

Conclusion: The paper demonstrates that without an intelligent, MEC-based optimization framework, DIC systems will fail to meet user needs during peak demand. The proposed MPC approach successfully transforms a constrained resource into a fair, efficient, and user-centric charging network.