# An Online Learning Framework for Energy-Efficient Navigation of Electric Vehicles

@inproceedings{kerblom2020AnOL, title={An Online Learning Framework for Energy-Efficient Navigation of Electric Vehicles}, author={Niklas {\AA}kerblom and Yuxin Chen and Morteza Haghir Chehreghani}, booktitle={IJCAI}, year={2020} }

Energy-efficient navigation constitutes an important challenge in electric vehicles, due to their limited battery capacity. We employ a Bayesian approach to model the energy consumption at road segments for efficient navigation. In order to learn the model parameters, we develop an online learning framework and investigate several exploration strategies such as Thompson Sampling and Upper Confidence Bound. We then extend our online learning framework to multi-agent setting, where multiple…

## 8 Citations

### A unified framework for online trip destination prediction

- Computer ScienceMachine Learning
- 2022

A unified framework for trip destination prediction in an online setting, suitable for both online training and online prediction, is presented and a novel regret metric is proposed for evaluating the entire online framework in comparison to its offline counterpart.

### Stochastic Route Planning for Electric Vehicles

- Computer ScienceSEA
- 2022

This work bridges the gap by finding routes to maximize on-time arrival probability and the set of non-dominated routes under two criteria for stochastic route feasibility: E -feasibility and p -Feasibility, and extends multi-criteria Contraction Hierarchies to accept Stochastic edge weights and offer heuristics to speed up queries.

### A Contextual Combinatorial Semi-Bandit Approach to Network Bottleneck Identification

- Computer ScienceArXiv
- 2022

This work develops a online learning framework based on combinatorial semi-bandits that performs bottleneck identiﬁcation alongside learning the speci⬁ca-tions of the underlying network.

### Passive and Active Learning of Driver Behavior from Electric Vehicles

- Computer Science2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC)
- 2022

This work investigates non-recurrent architectures such as self-attention models and convolutional neural networks with joint recurrence plots, and compares them with recurrent models, and demonstrates that some active sampling techniques can outperform random sampling, and therefore decrease the effort for annotation.

### Online Learning of Network Bottlenecks via Minimax Paths

- Computer ScienceMachine Learning
- 2022

This paper models the bottleneck identification in networks via extracting minimax paths as a combinatorial semi-bandit problem to which a combinatorsial version of Thompson Sampling is applied and establishes an upper bound on the corresponding Bayesian regret.

### Trip Prediction by Leveraging Trip Histories from Neighboring Users

- Computer Science2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC)
- 2022

It is shown empirical evidence that by enriching the users' trip histories with additional trips, one can improve the prediction error by 15%∼40%, evaluated on multiple subsets of the Nancy2012 dataset.

### Online Learning for Energy Efficient Navigation in Stochastic Transport Networks

- Computer Science
- 2021

This thesis introduces an online machine learning framework for navigation of electric vehicles, with the objective of adaptively and efficiently navigating the vehicle in a stochastic traffic environment and derives upper bounds on the Bayesian regret of Thompson Sampling adapted to this problem.

### Online Learning of Energy Consumption for Navigation of Electric Vehicles

- Computer ScienceArXiv
- 2021

An online learning framework is developed and several exploration strategies such as Thompson Sampling and Upper Conﬁdence Bound are investigated in a multi-agent setting, where multiple vehicles adaptively navigate and learn the parameters of the energy model.

## References

SHOWING 1-10 OF 34 REFERENCES

### Consumption Profiles in Route Planning for Electric Vehicles: Theory and Applications

- Computer ScienceSEA
- 2017

This work shows that the complexity of a profile is at most linear in the graph size, and derives a polynomial-time algorithm for the problem of finding an energy-optimal path between two locations that allows stops at charging stations.

### Energy consumption estimation integrated into the Electric Vehicle Routing Problem

- EngineeringTransportation Research Part D: Transport and Environment
- 2019

### The Shortest Path Problem Revisited: Optimal Routing for Electric Vehicles

- Computer ScienceKI
- 2010

This paper formalizes energy-efficient routing in the presence of rechargeable batteries as a special case of the constrained shortest path problem (CSPP) with hard and soft constraints, and presents an adaption of a general shortest path algorithm that respects the given constraints and has a worst case complexity of O(n3).

### (More) Efficient Reinforcement Learning via Posterior Sampling

- Computer ScienceNIPS
- 2013

An O(τS/√AT) bound on expected regret is established, one of the first for an algorithm not based on optimism, and close to the state of the art for any reinforcement learning algorithm.

### Efficient Energy-Optimal Routing for Electric Vehicles

- Computer ScienceAAAI
- 2011

This paper shows how battery constraints can be treated by dynamically adapting edge costs and hence can be handled in the same way as parameters given at query time, without increasing run-time complexity.

### Efficient Online Learning for Optimizing Value of Information: Theory and Application to Interactive Troubleshooting

- Computer ScienceUAI
- 2017

This work considers the optimal value of information problem, and proposes a dynamic hypothesis enumeration strategy, which allows efficient information gathering with strong theoretical guarantees and shows that with sufficient amount of samples, one can identify a near-optimal decision with high probability.

### Why is Posterior Sampling Better than Optimism for Reinforcement Learning?

- Computer ScienceICML
- 2017

An Bayesian expected regret bound for PSRL in finite-horizon episodic Markov decision processes is established, which improves upon the best previous bound of $\tilde{O}(H S \sqrt{AT})$ for any reinforcement learning algorithm.

### On Bayesian index policies for sequential resource allocation

- Computer Science, Mathematics
- 2016

It is proved that the Bayes-UCB algorithm, which relies on quantiles of posterior distributions, is asymptotically optimal when the reward distributions belong to a one-dimensional exponential family, for a large class of prior distributions.

### Luxembourg SUMO Traffic (LuST) Scenario: Traffic Demand Evaluation

- Computer ScienceIEEE Intelligent Transportation Systems Magazine
- 2017

The process used to build the Luxembourg SUMO Traffic (LuST) Scenario is shown, and a summary of its characteristics is presented together with the evaluation and validation of the traffic demand and mobility patterns.