Fair and Efficient Distributed EV Charging Coordination Using Threshold Cryptography and Machine Learning
Authors: Md. Mainul Islam, Muhammad Ismail, Hasan Kurban, Abdulrahman Takiddin, Erchin Serpedin
Organizations: Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77843, USA · Cybersecurity Education, Research, and Outreach Center (CEROC) and Department of Computer Science, Tennessee Tech University, Cookeville, TN 38505, USA · College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar · Department of Electrical and Computer Engineering, FAMU-FSU College of Engineering, Florida State University, Tallahassee, FL 32310 USA
Distributed coordination of electric vehicle (EV) charging with the alternating direction method of multipliers (ADMM) keeps EV constraints private but assumes that every participant follows the algorithm. This assumption fails in practice, where a biased operator or a man-in-the-middle (MITM) attacker can send inconsistent signals to different EVs, and compromised EVs can deviate from their prescribed updates. We first prove that attacks by such arbitrarily malicious, i.e., Byzantine, participants do not prevent convergence and therefore remain undetected, and derive the most damaging attack in closed form. We then propose a framework that combines a threshold-signed broadcast with upload recomputation, which detects every deviation without false alarms at linear instead of quadratic communication cost. To keep the optimization running, each affected EV recovers locally with a certified, fleet-aware machine learning price forecast, and we prove that the optimization remains convergent with a recovery cost bounded by the shape of the forecast error. Simulation results with real-world load, price, and charging data show that a biased operator raises the annual cost of disadvantaged EVs by 132.7% and lowers that of advantaged EVs by 49.3% without changing the aggregate load, whereas the defense limits all cost changes to 2.6%. An external MITM attacker raises the cost of its victims by 57.6%, the peak load by 5.8%, and the grid cost by 15.8%, whereas the defense keeps them within 1.1%, 0.2%, and 0.3% of benign operation. With up to 70% of the EVs attacked, the recovery converges in at least 97.8% of the runs.
Figures & tables
Work
Byzantine parties
Stealth a
Detection
Trusted channel b
Liveness c
Guarantees
Messages d
Attacks by participating EVs [ 15 , 16 ]
EVs
Yes
None
N/A
No
Attack analysis
N/A
Detection of deviating aggregators [ 17 ]
Aggregators
Partly
Statistical
N/A
No
Empirical
N/A
Byzantine-robust gradient aggregation [ 18 , 19 ]
Workers
No
Filtering
No
Partly
Statistical error bound
O(N)
Resilient distributed optimization [ 20 , 21 ]
Neighbors
No
Filtering
Yes (P2P)
Partly
Approximate convergence
O(L)
Trust-based node isolation [ 22 ]
EVs, links
Partly
Threshold-based
Yes (P2P)
Partly
Convergence after isolation
O(L)
Blockchain-based coordination [ 23 ]
Coordinator, participants
No
Consensus
Yes (ledger)
No
Consensus
O(N2)
TABLE I: Comparison With Prior Work on Attacks in Distributed Optimization
Fig. 1: Charging station and non-EV load served by a common bus.
Fig. 2: Framework architecture and adversary positions.
Fig. 3: Message flow of one certified iteration.
Model
Training
Testing
R2
MAE
R2
MAE
MAPE (%)
Span ($/kWh)
LightGBM
0.9964
0.0047
0.9695
0.0139
6.20
0.0550
Random forest
0.9926
0.0055
0.9666
0.0140
6.32
0.0552
XGBoost
0.9959
0.0054
0.9678
0.0145
6.59
0.0570
SVR
0.9638
0.0135
0.9699
0.0150
6.60
0.0581
Gradient boosting
0.9907
0.0083
0.9683
0.0151
6.81
0.0589
TABLE II: Training and Test Performance of the Price Forecasters
Fig. 4: Uncoordinated versus coordinated charging on a high-demand day: (a) total bus load and (b) EV load. The red band covers 5%–95% of 200 random uncoordinated draws, and gray marks the peak-price hours.
Fig. 5: Load and price profiles: (a) load under A1, (b) load under A2, and (c) final prices under A1 and A2. The shaded hours carry the peak price.
Fig. 6: Annual charging cost per EV under (a) A1 and (b) A2.
Fig. 7: Charging schedules, with daily cost changes in parentheses: (a) benign, (b) attacked, and (c) recovered under A1; (d)–(f) the same under A2.
Metric
Benign
A1: Biased operator
A2: MITM attacker
Undefended
Defended
Undefended
Defended
Change of J (%)
0.0
−0.2
−0.1
15.8
−0.3
LFS (%)
3.9
5.6
4.0
−1.6
4.1
Peak load change (%)
0.0
−1.7
−0.1
5.8
−0.1
Total EV payments ($)
928.6
926.5
927.4
1113.3
925.7
Station profit ($)
43.9
43.9
43.9
49.8
43.9
TABLE III: Case-Study Results Without and With the Proposed Defense
Fig. 8: Comparison of recovery rules over 77 held-out days: (a) convergence rate and (b) cost change of the attacked EVs. Markers show means, and bars show standard deviations over the days.
Fig. 9: Scalability of the defense: (a) iterations to convergence over all 365 daily instances, (b) latency of a certified iteration, and (c) messages per iteration.
Quantity
Plain ADMM
Proposed
P2P echo
Messages
2N=284
4N=568
2N2+N=40,470
Payload per EV (B)
384
722
384(2N+1)
Data volume (kB)
54.5
102.5
15,540
EV computation
4.3 μ s
2.5 ms
N checks
Operator computation
17.0 μ s
12.8 ms
–
Latency per iteration
0.021 ms
15.4 ms
–
TABLE IV: Computation and Communication Cost per Iteration ( N=142 )
College of Electronic and Information Engineering, Southwest University · Department of Electrical Engineering, City University of Hong Kong, Hong Kong · Department of Electrical and Electronic Engineering, the University of Hong Kong, Hong Kong