Trajectory-aware Dynamic Offloading for High-mobility Vehicular Edge Systems

doi:10.16182/j.issn1004731x.joss.25-0844

Abstract

Abstract:

To address the problems of service interruptions, task failures, and resource waste caused by high mobility of intelligent vehicles (IVs) in vehicular edge computing (VEC), this paper proposes a trajectory-aware dynamic offloading (TADO) framework for high-mobility vehicular edge systems. A lightweight T-pattern trajectory-aware algorithm is designed to efficiently predict the next-hop road side unit (RSU) by mining spatio-temporal patterns from historical trajectories of vehicles, offering a forward-looking reference for offloading decisions. A joint optimization model is constructed, and a service interruption risk factor driven by trajectory prediction is introduced. An improved DRL method is developed. It takes the predicted trajectories as key state inputs and integrates parameter space noise exploration with a prioritized experience replay mechanism to learn forward-looking and robust offloading strategies in highly dynamic environments. Simulation experiments based on real-world datasets of urban vehicle trajectories demonstrate that, compared to benchmark methods, such as local computing and full edge offloading, the TADO framework significantly reduces task processing latency and system energy consumption and maintains a high task completion rate across various scenarios.

Key words: vehicular edge computing, trajectory prediction, computation offloading, DRL

CLC Number:

TP393

Xu Haoran, Huang Zhiqin, Hu Youwu, Chen Zheyi. Trajectory-aware Dynamic Offloading for High-mobility Vehicular Edge Systems[J]. Journal of System Simulation, 2026, 38(7): 1935-1949.

Figures/Tables 12

Fig. 1

Fig. 2

Table 1

Settings of simulation parameters

仿真参数	设置
车辆 $v i$ 的传输功率 $P i$ /W	[0.4,0.6]^[18]
车辆 $v i$ 的本地计算能力 $f i l o c$ / $G H z$	[1.0,2.0] ^[19]
任务输入数据 $d i$ /MB	[0.4,1.2] ^[5]
任务计算复杂度 $c i$ /( $c y c l e s / b i t$ )	[200,1 000]
最大容忍延迟 $t i m a x$ / $s$	[1.5,3.0] ^[5,23]
高斯噪声功率 $σ 2$ /W	10^‒9[21]
单位距离信道增益 $g 0$ /dB	-30^{[18, 21]}
$λ 1$	0.7^[19]
$λ 2$	0.2^[19]
$λ 3$	0.1
经验回放池R	256
优先级权重 $ε$	0.6
重要性采样权重指数 $β$	0.4

Table 1

Fig. 3

Table 2

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Table 3

Fig. 8

Fig. 9

References 33

[1]	Li Haofei, Chen Chen, Shan Hangguan, et al. Deep Deterministic Policy Gradient-based Algorithm for Computation Offloading in IoV[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(3): 2522-2533.
[2]	Dai Xingxia, Xiao Zhu, Jiang Hongbo, et al. UAV-assisted Task Offloading in Vehicular Edge Computing Networks[J]. IEEE Transactions on Mobile Computing, 2024, 23(4): 2520-2534.
[3]	Zhang Haijun, Feng Lizhe, Liu Xiangnan, et al. User Scheduling and Task Offloading in Multi-tier Computing 6G Vehicular Network[J]. IEEE Journal on Selected Areas in Communications, 2023, 41(2): 446-456.
[4]	黄智钦, 卢恬英, 陈哲毅. 面向大规模IoT系统的多无人机部署与协作卸载[J]. 系统仿真学报, 2025, 37(1): 25-39.
	Huang Zhiqin, Lu Tianying, Chen Zheyi. Multi-UAV Deployment and Collaborative Offloading for Large-scale IoT Systems[J]. Journal of System Simulation, 2025, 37(1): 25-39.
[5]	Tian Shujuan, Zhu Xinjie, Feng Bochao, et al. Partial Offloading Strategy Based on Deep Reinforcement Learning in the Internet of Vehicles[J]. IEEE Transactions on Mobile Computing, 2025, 24(7): 6517-6531.
[6]	Fan Wenhao, Zhang Yaoyin, Zhou Guangtao, et al. Deep Reinforcement Learning-based Task Offloading for Vehicular Edge Computing with Flexible RSU-RSU Cooperation[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(7): 7712-7725.
[7]	Zeng Jiachen, Gou Fangfang, Wu Jia. Task Offloading Scheme Combining Deep Reinforcement Learning and Convolutional Neural Networks for Vehicle Trajectory Prediction in Smart Cities[J]. Computer Communications, 2023, 208: 29-43.
[8]	Wu Chaolun, Chiu Te-Chuan, Wang Chih-Yu, et al. Mobility-aware Deep Reinforcement Learning with Seq2seq Mobility Prediction for Offloading and Allocation in Edge Computing[J]. IEEE Transactions on Mobile Computing, 2024, 23(6): 6803-6819.
[9]	Xiao Hui, Hu Zhigang, Xu Aikun, et al. Cost and QoS-aware Service Migration in Vehicular Edge Computing Networks with Limited Backhaul Capacity[C]//2024 IEEE Smart World Congress (SWC). Piscataway: IEEE, 2024: 369-376.
[10]	Peng Yan, Tang Xiaogang, Zhou Yiqing, et al. Computing and Communication Cost-aware Service Migration Enabled by Transfer Reinforcement Learning for Dynamic Vehicular Edge Computing Networks[J]. IEEE Transactions on Mobile Computing, 2024, 23(1): 257-269.
[11]	Gao Zhen, Yang Lei, Dai Yu. VRCCS-AC: Reinforcement Learning for Service Migration in Vehicular Edge Computing Systems[J]. IEEE Transactions on Services Computing, 2024, 17(6): 4436-4450.
[12]	Chen Zheyi, Huang Sijin, Min Geyong, et al. Mobility-aware Seamless Service Migration and Resource Allocation in Multi-edge IoV Systems[J]. IEEE Transactions on Mobile Computing, 2025, 24(7): 6315-6332.
[13]	Ma Guifu, Wang Xiaowei, Hu Manjiang, et al. DRL-based Computation Offloading with Queue Stability for Vehicular-cloud-assisted Mobile Edge Computing Systems[J]. IEEE Transactions on Intelligent Vehicles, 2023, 8(4): 2797-2809.
[14]	Wu Xiang, Dong Jian, Bao Wei, et al. Augmented Intelligence of Things for Emergency Vehicle Secure Trajectory Prediction and Task Offloading[J]. IEEE Internet of Things Journal, 2024, 11(22): 36030-36043.
[15]	Zhang Xinyi, Wang Chunyang, Zhu Yanmin, et al. Multi-agent Deep Reinforcement Learning with Trajectory Prediction for Task Migration-assisted Computation Offloading[J]. IEEE Transactions on Mobile Computing, 2025, 24(7): 5839-5856.
[16]	Pazho A D, Noghre G A, Katariya V, et al. VT-former: An Exploratory Study on Vehicle Trajectory Prediction for Highway Surveillance Through Graph Isomorphism and Transformer[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Piscataway: IEEE, 2024: 5651-5662.
[17]	Xu Yunheng, Chen Jie, Wang Shuoheng, et al. TrajGNAS: Heterogeneous Multiagent Trajectory Prediction Based on a Graph Neural Architecture Search[C]//2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Piscataway: IEEE, 2025: 2532-2541.
[18]	Cheng Chen, Zhai Linbo, Zhu Xiumin, et al. Dynamic Task Offloading and Service Caching Based on Game Theory in Vehicular Edge Computing Networks[J]. Computer Communications, 2024, 224: 29-41.
[19]	Zhao Liang, Li Tianyu, Zhang Enchao, et al. Adaptive Swarm Intelligent Offloading Based on Digital Twin-assisted Prediction in VEC[J]. IEEE Transactions on Mobile Computing, 2024, 23(8): 8158-8174.
[20]	Sutton R S, Barto A. Reinforcement Learning: An Introduction[M]. 2nd ed. Cambridge: The MIT Press, 2020.
[21]	Wang Lei, Liang Hongbin, Zhao Dongmei. Deep-reinforcement-learning-based Computation Offloading and Power Allocation Within Dynamic Platoon Network[J]. IEEE Internet of Things Journal, 2024, 11(6): 10500-10512.
[22]	Huang Jiwei, Wan Jiangyuan, Bofeng Lü, et al. Joint Computation Offloading and Resource Allocation for Edge-cloud Collaboration in Internet of Vehicles via Deep Reinforcement Learning[J]. IEEE Systems Journal, 2023, 17(2): 2500-2511.
[23]	Wu Sijun, Yang Liang, Li Junjie, et al. DRL-based Pricing-driven for Task Offloading and Dynamic Resource in Vehicle Edge Computing[J]. IEEE Transactions on Mobile Computing, 2025, 24(10): 10389-10404.
[24]	Yan Ming, Xiong Rui, Wang Yan, et al. Edge Computing Task Offloading Optimization for a UAV-assisted Internet of Vehicles via Deep Reinforcement Learning[J]. IEEE Transactions on Vehicular Technology, 2024, 73(4): 5647-5658.
[25]	Zeng Jiaan, Kuang Zhufang, Chen Rui, et al. Delay-sensitive Dependent Tasks Offloading and Resource Allocation in VEC: A Deep-reinforcement Learning Approach[J]. IEEE Internet of Things Journal, 2025, 12(16): 34190-34203.
[26]	Ahmed Manzoor, Muhammad Ayzed Mirza, Raza Salman, et al. Vehicular Communication Network Enabled CAV Data Offloading: A Review[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(8): 7869-7897.
[27]	Gao Qinghang, Xiao Jianmao, Feng Zhiyong, et al. Optimization of Models and Strategies for Computation Offloading in the Internet of Vehicles: Efficiency and Trust[J]. IEEE Transactions on Mobile Computing, 2025, 24(4): 3372-3389.
[28]	Zhao Liang, Zhang Enchao, Wan Shaohua, et al. MESON: A Mobility-aware Dependent Task Offloading Scheme for Urban Vehicular Edge Computing[J]. IEEE Transactions on Mobile Computing, 2024, 23(5): 4259-4272.
[29]	Xiao Zhu, Xu Shenyuan, Li Tao, et al. On Extracting Regular Travel Behavior of Private Cars Based on Trajectory Data Analysis[J]. IEEE Transactions on Vehicular Technology, 2020, 69(12): 14537-14549.
[30]	Monreale Anna, Pinelli Fabio, Trasarti Roberto, et al. WhereNext: A Location Predictor on Trajectory Pattern Mining[C]//Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2009: 637-646.
[31]	Yao Liang, Xu Xiaolong, Bilal Muhammad, et al. Dynamic Edge Computation Offloading for Internet of Vehicles with Deep Reinforcement Learning[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(11): 12991-12999.
[32]	Jiang Hongbo, Cai Jianghao, Xiao Zhu, et al. Vehicle-assisted Service Caching for Task Offloading in Vehicular Edge Computing[J]. IEEE Transactions on Mobile Computing, 2025, 24(7): 6688-6700.
[33]	Piorkowski Michal, Sarafijanovic-Djukic Natasa, Grossglauser Matthias. CRAWDAD epfl/mobility[EB/OL]. (2022-11-15) [2025-07-25]. .

方法	精度/%	时间开销/ms
T-pattern	78.42	0.12
LSTM	84.60	12.45
Transformer	86.50	16.78

方法	任务最大容忍延迟/s
方法	1.5	2.0	3.0
TADO	62.31±1.12	78.80±0.95	92.68±0.82
DDPG	59.26±1.45	71.10±1.33	87.32±1.21
GA	55.96±1.87	69.76±1.56	85.71±1.54
EO	47.54±0.55	58.99±0.48	82.45±0.42
RO	49.12±0.63	60.34±0.59	79.20±0.51
LO	45.92±0	55.13±0	78.82±0