Typical mobile agent networks, such as multi-UAV systems, are constrained by limited resources: energy, computing power, memory and communication bandwidth. In particular, limited energy affects system performance directly, such as system lifetime. Moreover, it has been demonstrated experimentally in the wireless sensor network literature that the total energy consumption is often dominated by the communication cost, i.e. the computational and the sensing energy are small compared to the communication energy consumption. For this reason, the lifetime of the network can be extended significantly by minimizing the communication distance as well as the amount of communication data, at the expense of increasing computational cost. In this work, we aim at attaining an optimal trade-off between the communication and the computational energy. Specifically, we propose a mixed-integer optimization formulation for a multihop hierarchical clustering-based self-organizing UAV network incorporating data aggregation, to obtain an energy-efficient information routing scheme. The proposed framework is tested on two applications, namely target tracking and area mapping. Based on simulation results, our method can significantly save energy compared to a baseline strategy, where there is no data aggregation and clustering scheme.

Low cost unmanned aerial vehicles (UAVs) need multiple refuels to accomplish large area coverage. We propose the use of a mobile Ground Vehicle (GV), constrained to travel on a given road network, as a refueling station for the UAV. Determining optimal routes for a UAV and GV, and selecting rendezvous locations for refueling to minimize coverage time is NP-hard. We develop a two-stage strategy for coupled route planning for UAV and GV to perform a coverage mission. The first stage computes refueling sites that ensure reachability of all points of interest by the UAV and feasible routes for both the UAV and GV. In the second stage, Mixed-Integer Linear Programming (MILP)-based exact methods are developed to plan optimal routes for the UAV and GV. As the problem is NP-Hard, we also develop computationally efficient heuristics that can find good feasible solutions within a given time limit. Extensive simulations are conducted to corroborate the effectiveness of the developed approaches. Field experiments are also performed to verify the performance of the UAV-GV solution.

Fixed-wing unmanned aerial vehicles (UAVs) can be an essential tool for low-cost aerial surveillance and mapping applications in remote regions. There is, however, a key limitation, which is the fact that low-cost UAVs have limited fuel capacity and, hence, require periodic refueling to accomplish a mission. Moreover, the usual mechanism of commanding the UAV to return to a stationary base station for refueling can result in the fuel wastage and inefficient mission operation time. Alternatively, one strategy could be the use of an unmanned ground vehicle (UGV) as a mobile refueling unit, where the UAV will rendezvous with the UGV for refueling. In order to accurately perform this task in the presence of wind disturbances, we need to determine an optimal trajectory in three-dimensional taking UAV and UGV dynamics and kinematics into account. In this paper, we propose an optimal control formulation to generate a tunable UAV trajectory for rendezvous on a moving UGV that also addresses the possibility of the presence of wind disturbances. By a suitable choice of the value of an aggressiveness index that we introduce in our problem setting, we are able to control the UAV rendezvous behavior. Several numerical results are presented to illustrate the reliability and effectiveness of our approach.

Autonomous landing on a moving target is challenging because of external disturbances and localization errors. In this paper, we present a vision-based guidance technique with a log polynomial closing velocity controller to achieve faster and more accurate landing as compared to that of the traditional vertical landing approaches. The vision system uses a combination of color segmentation and AprilTags to detect the landing pad. No prior information about the landing target is needed. The guidance is based on pure pursuit guidance law. The convergence of the closing velocity controller is shown, and we test the efficacy of the proposed approach through simulations and field experiments. The landing target during the field experiments was manually dragged with a maximum speed of $0.6$ m/s. In the simulations, the maximum target speed of the ground vehicle was 3 m/s. We conducted a total of 27 field experiment runs for landing on a moving target and achieved a successful landing in 22 cases. The maximum error magnitude for successful landing was recorded to be 35 cm from the landing target center. For the failure cases, the maximum distance of vehicle landing position from target boundary was 60 cm.

Determining multi-robot motion policies for persistently monitoring a region with limited sensing, communication, and localization constraints in non-GPS environments is a challenging problem. To take the localization constraints into account, in this paper, we consider a heterogeneous robotic system consisting of two types of agents: anchor agents with accurate localization capability and auxiliary agents with low localization accuracy. To localize itself, the auxiliary agents must be within the communication range of an anchor, directly or indirectly. The robotic team’s objective is to minimize environmental uncertainty through persistent monitoring. We propose a multi-agent deep reinforcement learning (MARL) based architecture with graph convolution called Graph Localized Proximal Policy Optimization (GALOPP), which incorporates the limited sensor field-of-view, communication, and localization constraints of the agents along with persistent monitoring objectives to determine motion policies for each agent. We evaluate the performance of GALOPP on open maps with obstacles having a different number of anchor and auxiliary agents. We further study 1) the effect of communication range, obstacle density, and sensing range on the performance and 2) compare the performance of GALOPP with area partition, greedy search, random search, and random search with communication constraint strategies. For its generalization capability, we also evaluated GALOPP in two different environments – 2-room and 4-room. The results show that GALOPP learns the policies and monitors the area well. As a proof-of-concept, we perform hardware experiments to demonstrate the performance of GALOPP. Note to Practitioners —Persistent monitoring is performed in various applications like search and rescue, border patrol, wildlife monitoring, etc. Typically, these applications are large-scale, and hence using a multi-robot system helps achieve the mission objectives effectively. Often, the robots are subject to limited sensing range and communication range, and they may need to operate in GPS-denied areas. In such scenarios, developing motion planning policies for the robots is difficult. Due to the lack of GPS, alternative localization mechanisms, like SLAM, high-accurate INS, UWB radio, etc. are essential. Having SLAM or a highly accurate INS system is expensive, and hence we use agents having a combination of expensive, accurate localization systems (anchor agents) and low-cost INS systems (auxiliary agents) whose localization can be made accurate using cooperative localization techniques. To determine efficient motion policies, we use a multi-agent deep reinforcement learning technique (GALOPP) that takes the heterogeneity in the vehicle localization capability, limited sensing, and communication constraints into account. GALOPP is evaluated using simulations and compared with baselines like random search, random search with ensured communication, greedy search, and area partitioning. The results show that GALOPP outperforms the baselines. The GALOPP approach offers a generic solution that be adopted with various other applications.

Persistent monitoring on terrains using mobile robotic sensors requires coordinated planning. Terrain features add visibility obstacles and limited fuel capacity of aerial robots leads to range restrictions that make the problem challenging. We address the visual-monitoring problem on piecewise linear features within a terrain using multiple mobile robots for persistent operations. The planner must account for visual coverage, refueling aerial robots during the mission, and placement of refueling depots while also utilizing the available sensor diversity to minimize overall costs for the monitoring mission. Building on previous works on visibility in specific classes of polygons and fuel-constrained routing, we develop a discrete representation of the problem that allows the design and application of discrete optimization techniques to find optimal solutions. We develop a mixed-integer linear programming (MILP) formulation and discuss a branch-and-cut implementation to compute exact solutions. We also develop a construction heuristic based on the idea of competitive construction of robot paths using a step-increment strategy. We report the results from computational simulations and illustrate proof of concept using experiments on real robots. Note to Practitioners —This article is motivated by the need to perform persistent monitoring in applications, such as border patrol and perimeter surveillance. Unmanned aerial and ground robots can be used to perform these activities uninterruptedly. However, aerial robots have limited fuel capacity and need periodic refueling. Hence, the number of refueling depots and their placement within the environment also affects the monitoring task. Also, due to terrain variation, robots are subject to limited visibility. Therefore, we need to consider refueling constraints and terrain visibility aspects while planning optimal routes for the robots to perform visual monitoring. In this article, we present a general optimal routing formulation to compute exact solutions. We also present a fast heuristic for real-time applications that produce feasible solutions. The algorithms are validated in simulations. We also show a proof of concept using experiments in limited outdoor settings.