Zhiyuan Lin

Zhiyuan Lin

Auto-generated profile, based on publicly available information

Zhiyuan Lin

Tsinghua University

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 January 2022 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
ZLZhiyuan Lin
Zhiyuan Lin
ZNZuyao Ni
Zuyao Ni
LKLinling Kuang
Linling Kuang
CJChunxiao Jiang
Chunxiao Jiang
ZHZhen Huang
Zhen Huang
Due to the non-uniform geographic distribution and time-varying characteristics of the ground traffic request, how to make full use of the limited beam resources to serve users flexibly and efficiently is a brand-new challenge for beam hopping satellite systems. The conventional greedy-based beam hopping methods do not consider the long-term reward, which is difficult to deal with the time-varying traffic demand. Meanwhile, the heuristic algorithms such as genetic algorithm have a slow convergence time, which can not achieve real-time scheduling. Furthermore, existing methods based on deep reinforcement learning (DRL) only make decisions on beam patterns, lack of the freedom of bandwidth. This paper proposes a dynamic beam pattern and bandwidth allocation scheme based on DRL, which flexibly uses three degrees of freedom of time, space and frequency. Considering that the joint allocation of bandwidth and beam pattern will lead to an explosion of action space, a cooperative multi-agents deep reinforcement learning (MADRL) framework is presented in this paper, where each agent is only responsible for the illumination allocation or bandwidth allocation of one beam. The agents can learn to collaborate by sharing the same reward to achieve the common goal, which refers to maximize the throughput and minimize the delay fairness between cells. Simulation results demonstrate that the offline trained MADRL model can achieve real-time beam pattern and bandwidth allocation to match the non-uniform and time-varying traffic request. Furthermore, when the traffic demand increases, our model has a good generalization ability.

1 December 2022 in IEEE Transactions on Communications
Verified
Journal Article
Closed access
ZLZhiyuan Lin
Zhiyuan Lin
ZNZuyao Ni
Zuyao Ni
LKLinling Kuang
Linling Kuang
CJChunxiao Jiang
Chunxiao Jiang
ZHZhen Huang
Zhen Huang
Due to the non-uniform distribution of the ground traffic demand and the high mobility of non-geostationary orbit (NGSO) satellites, how to make full use of the limited beam resources to serve users flexibly and efficiently is a brand-new challenge for NGSO communication systems. In order to achieve efficient spectrum utilization, the combination of full frequency multiplexing and beam hopping is a major trend in future satellite communication systems. However, conventional beam hopping methods are mainly based on geostationary satellites, which do not take into account the interference between satellites. This paper proposes a multi-satellite beam hopping algorithm based on load balancing and interference avoidance, which takes advantage of the multiple coverage features in the NGSO constellation and avoids intra-satellite interference and inter-satellite interference by designing beam-hopping patterns with spatial isolation characteristics. In particular, we decompose the multi-satellite beam hopping problem into three sub-problems, which are the multi-satellite load balancing problem, the single-satellite beam hopping pattern design problem, and the multi-satellite interference avoidance problem. Simulation results demonstrate that the proposed method reduces the load gap among satellites by about 72.5% and the average traffic satisfaction rate can reach 81.4%. Besides, our method has the lowest unmet capacity compared with other benchmarks, achieving better offered-requested data match.

29 February 2024 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
ZLZhiyuan Lin
Zhiyuan Lin
ZNZuyao Ni
Zuyao Ni
LKLinling Kuang
Linling Kuang
CJChunxiao Jiang
Chunxiao Jiang
ZHZhen Huang
Zhen Huang
Non-geostationary orbit (NGSO) constellations enabled by beam hopping (BH) technology are characterized by wide coverage and high spectrum efficiency. However, how to efficiently schedule multi-satellite beam resources to satisfy the heterogeneous and uneven terrestrial traffic demands remains a huge challenge for satellite operators. This paper proposes a satellite-terrestrial coordinated multi-satellite BH scheduling framework, where the complex multi-satellite BH problem is formulated into a long-term and a short-term subproblems. The long-term subproblem is cell-satellite association problem, which is solved by a low-complexity iterative algorithm executed in network operation control center (NOCC) to minimize the traffic load gap among satellites while considering the interference avoidance. The short-term subproblem is multi-satellite traffic-driven BH problem and we propose a multi-agent deep reinforcement learning (MADRL) architecture where each satellite can cooperatively make real-time BH decisions using the well-trained model by QMIX algorithm to adapt to time-varying and heterogeneous traffic. Simulation results demonstrate that the traffic load gap and network delay have been reduced by 70% and 50% respectively compared with non-load-balancing scheme. Besides, the proposed algorithm outperforms other benchmarks in terms of the network throughput under various traffic load cases and the average network delay is kept within 4 ms. Furthermore, the proposed QMIX-BH can be applied to real-time scheduling since the execution time is less than 1 ms.

12 October 2022 in IEEE Communications Letters
Verified
Journal Article
Closed access
ZLZhiyuan Lin
Zhiyuan Lin
ZNZuyao Ni
Zuyao Ni
LKLinling Kuang
Linling Kuang
CJChunxiao Jiang
Chunxiao Jiang
ZHZhen Huang
Zhen Huang
Non-geostationary orbit (NGSO) satellite communication system plays an important role in achieving seamless global coverage and addressing the digital divide. The uneven distribution of terrestrial users and the high mobility of NGSO satellites put forward higher requirements for resource management. In this letter, we propose a multi-satellite beam hopping scheme based on load balancing and interference avoidance, where the complicated multi-satellite joint beam hopping problem is decomposed into two sub-problems, namely the satellite-cell assignment problem and joint multi-satellite beam hopping pattern design problem. On the one hand, we rationally assign cells to different satellites for balancing the traffic load among satellites and reducing the interference to geostationary earth orbit (GSO) ground stations. On the other hand, an iterative multi-satellite beam hopping pattern design scheme based on minimizing co-channel interference criterion is adopted to achieve on-demand service. Simulation results indicate that the performance of load balancing and interference avoidance are improved compared to the maximum elevation angle access strategy. Besides, the proposed scheme outperforms other benchmarks in terms of the traffic satisfaction rate, and the optimality gap is only 0.1%.

22 January 2025 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
ZLZhiyuan Lin
Zhiyuan Lin
LKLinling Kuang
Linling Kuang
BLBingkun Liu
Bingkun Liu
CJChunxiao Jiang
Chunxiao Jiang
JJJin Jin
Jin Jin
Non-geostationary orbit (NGSO) constellations characterized by seamless coverage and high throughput have been envisioned as a promising solution to 6G, where beam hopping (BH) technique plays an indispensable role to achieve on-demand service. However, owing to the dynamic topology and uneven terrestrial traffic, how to effectively manage multi-satellite time-frequency-spatial beam resources to match the heterogeneous demands is challenging for NGSO satellite operators. This paper proposes a novel return-link BH scheduling framework for NGSO mega-constellations, where the ground resource management center periodically executes BH scheduling for users in earth-fixed cells to improve the long-term average service satisfaction (ASS) while reducing the inter-satellite handover. Specifically, we decouple the NP-hard scheduling problem into three subproblems, namely user-satellite association problem, BH pattern design problem and resource block (RB) allocation problem. User-satellite association problem is solved by a low-complexity heuristic algorithm to reduce the handover and avoid inter-satellite interference. Furthermore, a graph-based maximum weight method and an iterative scheme based on successive convex approximation are developed to obtain the BH patterns and RB allocations, respectively. Simulation results demonstrate that compared with other benchmarks such as the best channel association and round-robin beam hopping schemes, the proposed method can reduce the satellite load unevenness by 60% and improve the ASS by 35% while maintaining a low handover overhead.

20 May 2026 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
RWRuiqing Wen
Ruiqing Wen
JJJin Jin
Jin Jin
ZLZhiyuan Lin
Zhiyuan Lin
LKLinling Kuang
Linling Kuang
In non-geostationary orbit (NGSO) satellite systems, beam hopping (BH) emerges as a promising technique, enabling satellites to efficiently address dynamic and unevenly distributed ground user demands with fewer beams, thereby optimizing resource utilization. However, the orbital dynamics of NGSO satellites cause continual variations in individual satellite service cells and inter-satellite coordination, challenging the real-time performance and generalization of BH algorithms. To tackle this, we propose a cooperative BH architecture leveraging the attention mechanism and multi-agent communication. Firstly, enhanced load balancing is achieved by solving the satellite-user association problem for greater flexibility. Secondly, building upon the centralized training with decentralized execution (CTDE) paradigm, we designed an attention-based network backbone for real-time single-satellite BH to adapt to dynamically changing cell sets. Moreover, a periodic multi-agent communication framework is proposed, allowing satellites to sustain coordination in evolving environments through information exchange over long-term operations. Simulation results show that the proposed algorithm outperforms other benchmark methods. Compared to multi-layer perceptron (MLP)-based multi-agent deep reinforcement learning (MADRL) algorithms, it yields a 17% throughput improvement under equivalent average user delay. Furthermore, we demonstrate the feasibility of inter-satellite multi-agent communication via inter-satellite links (ISLs), revealing that this enhancement improves performance while occupying less than 0.01% of ISL bandwidth per message, underscoring its practicality for real-world deployment.