Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Sunil Kumar has a total of 53 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Doppler-Adaptive Radio Waveform Engine for Highly Dynamic FANETs Using Time-Fused Transformer Reinforcement Learning

    by Linsheng He, Weiqiang Lyu, Fei Hu, Elizabeth Serena Bentley, Sunil Kumar

    Flying ad hoc networks (FANETs), consisting of unmanned aerial vehicles (UAV), and aircraft have become popular in the 5G era. However, a fast-response adaptive waveform control mechanism tailored for FANET applications has not yet been deeply investigated. In various scenarios, such as slow but crowded drone swarming networks for search and rescue, fast drone-show networks, and aircraft formations, it is a significant challenge to ensure that the RF modulation method can timely cope with the Doppler effect caused by node mobility changes. This paper introduces a Doppler adaptive waveform engine (DAWE) to adjust the node’s modulation method and parameters, according to the highly dynamic FANET conditions. An enhanced orthogonal time frequency space (OTFS) modulation method is implemented in DAWE to mitigate high Doppler shifts and multipath propagation across diverse FANET scenarios. We further propose a time-fused Transformer mechanism-based reinforcement learning (TTRL) algorithm to adjust the modulation parameters. This algorithm can focus on long-range dependencies better than traditional time-series forecast models and does not require the actor-critic dual model structure that is commonly used in deep reinforcement learning. Our simulations show that, across link SNRs of 20 – 40 dB and user speeds ranging from 100 km h¹ (pedestrian UAV loiter) up to 1000 km h¹ (high-speed fixed-wing transit), the DAWE protocol sustains a bit-error rate roughly one order of magnitude ( 10 ×) lower than a conventional orthogonal frequency-division multiplexing(OFDM)-only baseline.


  • A Hybrid Reactive Routing Protocol for Decentralized UAV Networks

    by Shivam Garg, Alexander Ihler, Elizabeth Serena Bentley, Sunil Kumar

    Wireless networks consisting of low SWaP (size, weight, and power), fixed-wing UAVs (unmanned aerial vehicles) are used in many applications, such as monitoring, search, and surveillance of inaccessible areas. A decentralized and autonomous approach ensures robustness to failures; the UAVs explore and sense within the area and forward their information, in a multihop manner, to nearby aerial gateway nodes. However, the unpredictable nature of the events, relatively high speed of the UAVs and dynamic trajectories cause the network topology to change significantly over time, resulting in frequent route breaks. A holistic routing approach is needed to support multiple traffic flows in these networks to provide mobility- and congestion-aware, high-quality routes when needed, with low control and computational overhead, using the information collected in a distributed manner. Existing routing schemes do not address all the mentioned issues. This paper presents a hybrid reactive routing protocol for decentralized UAV networks called Hyd-AODV. It searches routes on-demand (using a multi-metric route selection), monitors a region around the selected route (the “pipe”), and proactively switches to an alternative route before the current route's quality degrades below a threshold. The impact of pipe width is empirically and theoretically evaluated to find alternate high-quality routes within the pipe and the overhead required to maintain the pipe. A queue management scheme is also incorporated to prioritize packet transmissions based on their age of information (AoI). Compared to existing reactive routing schemes, the proposed approach achieves higher throughput and reduces the number of route discoveries, overhead, and resulting flow interruptions at different traffic loads, node density, and speeds. Despite having limited network topology information and low overhead and route computation complexity, the proposed scheme achieves a superior throughput to proactive optimized link state routing (OLSR) scheme for different network and traffic settings. The relative performance of reactive and proactive routing schemes is also studied.


  • Open Access

    Pipe Routing with Topology Control for Decentralized and Autonomous UAV Networks

    by Shreyas Devaraju, Shivam Garg, Alexander Ihler, Elizabeth Serena Bentley, Sunil Kumar

    This paper considers a decentralized and autonomous wireless network of low SWaP (size, weight, and power) fixed-wing UAVs (unmanned aerial vehicles) used for remote exploration and monitoring of targets in an inaccessible area lacking communication infrastructure. Here, the UAVs collaborate to find target(s) and use routing protocols to forward the sensed data of target(s) to an aerial base station (BS) in real-time through multihop communication, which can then transmit the data to a control center. However, the unpredictability of target locations and the highly dynamic nature of autonomous, decentralized UAV networks result in frequent route breaks or traffic disruptions. Traditional routing schemes cannot quickly adapt to dynamic UAV networks and can incur large control overhead and delays. In addition, their performance suffers from poor network connectivity in sparse networks with multiple objectives (exploration and monitoring of targets), which results in frequent route unavailability. To address these challenges, we propose two routing schemes: Pipe routing and TC-Pipe routing. Pipe routing is a mobility-, congestion-, and energy-aware scheme that discovers routes to the BS on-demand and proactively switches to alternate high-quality routes within a limited region around the routes (referred to as the “pipe”) when needed. TC-Pipe routing extends this approach by incorporating a decentralized topology control mechanism to help maintain robust connectivity in the pipe region around the routes, resulting in improved route stability and availability. The proposed schemes adopt a novel approach by integrating the topology control with routing protocol and mobility model, and rely only on local information in a distributed manner. Comprehensive evaluations under diverse network and traffic conditions—including UAV density and speed, number of targets, and fault tolerance—show that the proposed schemes improve throughput by reducing flow interruptions and packet drops caused by mobility, congestion, and node failures. At the same time, the impact on coverage performance (measured in terms of coverage and coverage fairness) is minimal, even with multiple targets. Additionally, the performance of both schemes degrades gracefully as the percentage of UAV failures in the network increases. Compared to schemes that use dedicated UAVs as relay nodes to establish a route to the BS when the UAV density is low, Pipe and TC-Pipe routing offer better coverage and connectivity trade-offs, with the TC-Pipe providing the best trade-off.


  • Open Access

    Directional Multicast Routing in Mobile Networks Through Tracking of Directional Interference Dynamics

    by Jiamiao Zhao, Linsheng He, Fei Hu, Sunil Kumar

    This research targets interference-resistant routing in airborne multicast networks with directional antennas (DAs). Multicast communications are important in one-to-many (1:M) message deliveries. Wireless multicast protocol design is more challenging than conventional wired multicast networks, due to the dynamic network topology caused by the frequent node mobility. On the other hand, the deploying of DAs in each node enhances the quality due to its well-focused wireless transmissions. However, when DA is equipped in each node of the multicast tree, the signal interference becomes more difficult to handle than in the case of omnidirectional networks, due to the special features of directional interference (DI), which varies with the parent-children spatial relationship in the multicast tree. The DI becomes more challenging when the multicast tree gets too large (such that it must be decomposed into multiple subtrees). Both intra- and inter-subtree DI cases exist, and different strategies are required to overcome the impacts of the interference. In this research, we propose a predictive multicast DI avoidance scheme, which adopts spatio-temporal deep learning with up/down sampling to perform both spatial and temporal DI evolution prediction. Such predicted DI patterns in intra-/inter-subtrees will be used to design a directional multicast cross-layer (routing/MAC) protocol. Such a cross-layer protocol design is further enhanced with a bio-inspired scheme, i.e., an enhanced ant colony optimization (ACO) algorithm. Our simulation results verify the efficiency of the DI-prediction-based cross-layer design in the directional multicast networks.


  • Directional Interference Avoidance With Minimum Interaircraft Interactions for Flying Ad Hoc Networks (FANET) With Directional Antennas Through Distributed Multiagent Reinforcement Learning

    by Linsheng He, Jiamiao Zhao, Fei Hu, Sunil Kumar

    Flying ad hoc networks (FANETs) communications have started to use next-generation wireless technologies, such as terahertz (THz) networks. Conventional protocols often require a central coordinator (the main aircraft) and frequent message exchanges among aircraft by using omnidirectional antennas to make optimal communication adjustments. Those schemes cannot utilize the advantages of directional antennas used in THz communications. Moreover, it could significantly diminish the efficiency of high-speed THz links. Hence, we introduce a distributed adaptive layer (DAL) in the protocol stack to adjust the node's attributes according to the dynamic network properties in FANETs. Here the attributes include the sending rate, antenna facing angle, antenna beamwidth, directional interference levels (defined in the concept of exclusive region (ER)), and other flying parameters. We have also implemented the Multi-Agent Deep Distributed Reinforcement Learning (MADDRL) algorithm within the DAL protocol. Our decentralized intelligence model requires minimal information exchange among one-hop neighbors and enables each node/agent to interact with its environment based on individual and local observations. Furthermore, we employ two primary communication evaluation metrics, i.e., the end-to-end throughput and queueing delay, to showcase the effectiveness of our approach. Based on our simulations, the DAL protocol exhibits several advantages over the conventional schemes that rely on a central controller. These include 150% higher throughput, nearly 4 times less delay, and the reduction of protocol message exchanges.


  • A Deep-$Q$-Learning-Based Base-Station-Connectivity-Aware Decentralized Pheromone Mobility Model for Autonomous UAV Networks

    by Shreyas Devaraju, Alexander Ihler, Sunil Kumar

    Wireless networks consisting of low SWaP (size, weight, and power), fixed-wing UAVs (unmanned aerial vehicles) are used in many applications, such as search, monitoring and information gathering of inaccessible areas, in which UAVs sense within an area and forward the information, in a multihop manner, to an aerial base station (BS). Robustly performing these tasks requires the UAV network to be decentralized, autonomous, and scalable. An important tradeoff is between area coverage and connectivity: fast area coverage is needed to quickly identify objects of interest, while connectivity must be maintained for coordination and to transmit sensed information to the BS in real time. These factors must be balanced by the mobility model, which for each UAV has access only to locally available information. While [1], [2] attempt to balance these factors using flocking behavior, this only encourages the UAVs to spread, rather than using knowledge of what areas have already been covered. In this paper, we develop a neighborhood- and BS-connectivity aware distributed pheromone mobility model, called BS-CAP, to autonomously coordinate the UAV movements in a decentralized network. By using a pheromone map, we directly incorporate recent coverage information for the area. We then extend our approach to a deep Q-learning policy variant, called BSCAPDQN, to further tune and improve the balance between coverage and connectivity. These mobility models are fully distributed and rely only on information from neighboring UAVs. Our simulations demonstrate that both models achieve efficient area coverage and improved connectivity (both locally and to the BS), providing significant improvements over existing approaches.


  • Open Access

    Pipe Routing with Topology Control for UAV Networks

    by Shreyas Devaraju, Shivam Garg, Alexander Ihler, Sunil Kumar

    Routing protocols help in transmitting the sensed data from UAVs monitoring the targets (called target UAVs) to the BS. However, the highly dynamic nature of an autonomous, decentralized UAV network leads to frequent route breaks or traffic disruptions. Traditional routing schemes cannot quickly adapt to dynamic UAV networks and/or incur large control overhead and delays. To establish stable, high-quality routes from target UAVs to the BS, we design a hybrid reactive routing scheme called pipe routing that is mobility, congestion, and energy-aware. The pipe routing scheme discovers routes on-demand and proactively switches to alternate high-quality routes within a limited region around the active routes (called the pipe) when needed, reducing the number of route breaks and increasing data throughput. We then design a novel topology control-based pipe routing scheme to maintain robust connectivity in the pipe region around the active routes, leading to improved route stability and increased throughput with minimal impact on the coverage performance of the UAV network.

  • Open Access

    A Deep Q-Learning based, Base-Station Connectivity-Aware, Decentralized Pheromone Mobility Model for Autonomous UAV Networks

    by Shreyas Devaraju, Alexander Ihler, Sunil Kumar

    UAV networks consisting of low SWaP (size, weight, and power), fixed-wing UAVs are used in many applications, including area monitoring, search and rescue, surveillance, and tracking. Performing these operations efficiently requires a scalable, decentralized, autonomous UAV network architecture with high network connectivity. Whereas fast area coverage is needed for quickly sensing the area, strong node degree and base station (BS) connectivity are needed for UAV control and coordination and for transmitting sensed information to the BS in real time. However, the area coverage and connectivity exhibit a fundamental trade-off: maintaining connectivity restricts the UAVs' ability to explore. In this paper, we first present a node degree and BS connectivity-aware distributed pheromone (BS-CAP) mobility model to autonomously coordinate the UAV movements in a decentralized UAV network. This model maintains a desired connectivity among 1-hop neighbors and to the BS while achieving fast area coverage. Next, we propose a deep Q-learning policy based BS-CAP model (BSCAP-DQN) to further tune and improve the coverage and connectivity trade-off. Since it is not practical to know the complete topology of such a network in real time, the proposed mobility models work online, are fully distributed, and rely on neighborhood information. Our simulations demonstrate that both proposed models achieve efficient area coverage and desired node degree and BS connectivity, improving significantly over existing schemes.

  • Intelligent Routing in Directional Ad Hoc Networks Through Predictive Directional Heat Map From Spatio-Temporal Deep Learning

    by Zhe Chu, Fei Hu, Elizabeth Bentley, Sunil Kumar

    By applying a simple shortest/minimum-cost routing algorithm, the mobile ad-hoc network (MANET) with heavy data transmissions may be easily congested if multiple routes meet at the same relay node. Therefore, those busy nodes should be avoided when a new path is established. The task of optimal path seeking becomes more challenging when a MANET is equipped with directional antennas that may cause directional interference with neighboring receivers. The motivation of our research is to build an intelligent proactive routing scheme for MANETs with directional antennas. Our directional routing protocol considers not only the global traffic distribution in different areas of the MANET, but also the properties of directional antennas. It uses a spatio-temporal deep learning algorithm to predict the next-time snapshot of a directional heat map (DHM), which shows the traffic density distribution in each network location as well as the coverage of each directional antenna. The DHM is then used to identify the optimal path that can avoid congested areas as well as the interference from all neighboring directional links. Furthermore, an optimization algorithm is designed to perform optimal path selection. It splits a single path into multiple paths converge later on into one path, if the path needs to go around a congested area. Therefore, our routing scheme achieves better quality-of-service (QoS) performance than existing routing schemes.


  • A Cross-Layer, Mobility, and Congestion-Aware Routing Protocol for UAV Networks

    by Shivam Garg, Alexander Ihler, Elizabeth Serena Bentley, Sunil Kumar

    Autonomous and decentralized UAV networks are used in many applications, including disaster management, environment monitoring and surveillance. The performance of the routing protocols in these networks drops significantly due to frequent route breaks and traffic congestion at high node speeds. This is partly because the routing metrics used in many of the existing protocols are sensitive to the intra-flow interference after the source node starts transmission on the selected route. Moreover, these schemes do not estimate potential packet drops or take preventive measures, such as route switching and queue management. To address these issues, we present a cross-layer, mobility and congestion-aware optimized link state routing (MCA-OLSR) protocol for an autonomous and decentralized UAV network, which incorporates the following mechanisms: ( i ) Multi-metric route selection which includes hop count, route lifetime, estimated time-to-destination and number of interfering links; ( ii ) Preemptive route switching which prevents packet transmission over the broken and congested routes; and ( iii ) Queue management to prioritize transmission of packets based on their survivability score. We also present key observations on routing protocol design and a theoretical analysis to show the impact of data rate, route length and interfering links on the packet service rate and congestion. The proposed routing scheme achieves a significantly higher packet delivery ratio and lower route computation overhead for delay-sensitive data flows at different data rates, node densities and high speeds, as compared to the standard and multi-metric OLSR protocols.