Particle-dynamics modeling of drone swarmsShabaev, A.; Chichka, D.; Ward, E.; Howells, C.; Lambrakos, S. G.
doi: 10.1117/12.2661331pmid: N/A
The concept of enabling drone-swarm engagement simulations using particle-dynamics models and near-neighbors tracking algorithms, motivated by SDI battle management, is examined. The general approach of using particle-dynamics models and near-neighbors tracking algorithms for modeling drone-swarm engagements is similar to nonequilibrium molecular-dynamics modeling of mixing dissimilar particulate materials. With respect to particle-dynamics representation of swarm-engagements, fundamental quantities that can represent characteristics of drone interactions, are interparticle potential functions, which are a function of drone-drone separation, the types of drones interacting, and the nature of the interaction. These potential functions provide formal representation of both deterministic and non-deterministic dronedrone interaction scenarios. The complexity of drone-swarm engagements, similar to that of SDI scenarios, characterized by small time-periods of engagement, multiple types of blue-red force interactions, and the requirement of near-neighbor target tracking, suggest that such a tool be necessary. The utility of the tool in creating potential-theory based control algorithms for swarm-on-swarm engagements is demonstrated using particle-dynamics simulations.
Path planning and obstacle avoidance utilizing chameleon swarm algorithmBallous, Khaled Awad; AlShabi, Mohammad; Bou Nassif, Ali; Bettayeb, Maamar
doi: 10.1117/12.2664081pmid: N/A
Path planning and obstacle avoidance are crucial tasks in the robotics and autonomous industry. Path planning seeks to determine the most efficient path between a start and an end point, whereas obstacle avoidance seeks to avoid collisions with static or dynamic obstacles in the environment. On this work, we utilize the Chameleon Swarm Algorithm (CSA), which is a metaheuristic approach, for path planning and obstacle avoidance on a predetermined map with static obstacles. This CSA extracted the optimal path from several possible different paths, and the results showed that it has slightly superior performance compared to PSO.
An intelligent mission-oriented and situation aware quadruped robot: a novel embodied explainable AI approachLokhande, Sanket; Dailey, Joseph; Liu, Yuqing; Connolly, Samantha; Xu, Hao
doi: 10.1117/12.2664019pmid: N/A
Quadruped locomotion and gait patterns as trotting, galloping, trot running has been investigated and applied to a variety of existing quadruped robots such as Big Dog from Boston Dynamics, A1 Robot dog from Unitree. Most of these studies are based either on biology inspired gaits or the best possible locomotion that can be performed by the individual robot with its pre-set mechanics and its availability of the degree of freedoms. While these are already available as their basic modes, a wide number of researchers are investigating locomotion via deep neural nets. These are making headlines in the research community for efficiency of use, and yet the explainability is lacking in most cases. Just like a Large Language Model might give spurious results here and there for basic common sense questions, these deep neural nets also make errors with unknown interpretability to the inputs. Regarding training, they require careful tuning of hyperparameters and training with a number of parameters unknown to user predictions. For example, on the field we might have a terrain which is flat for a certain length, in addition to a rocky climb, followed by a slippery slope. The combinations are as many as possible and the existing state of the art is heavily depending on human intervention and training predictions to handle the change of modes of the gait patterns that can fit into the terrain underneath. In this paper, we develop a novel embodied explainable machine learning algorithm which can help minimize the training as well as human intervention when autonomous operations are required. Specifically, we utilize the Markov Decision Process (MDP) along with rules set forth by DARPA in the Explainable AI (XAI) research. The XAI research enables us to generate textual explanations of the behavior by utilizing the MDP and reinforcement learning to generate mission oriented and situation aware cost functions along with the ones which are already pre-programmed. We validate our hypothesis in real hardware across different conditions.
Thermal modeling for autonomous vehicle simulationsCecil, Orie M.; Carrillo, Justin T.; Bray, Matthew D.; Trautz, Andrew C.; Brady, Brian R.; Monroe, John G.; Farthing, Matthew W.; Fairley, Josh R.
doi: 10.1117/12.2663842pmid: N/A
Autonomous vehicles (AVs) employ a wide range of sensing modalities including LiDAR, radar, RGB cameras, and more recently infrared (IR) sensors. IR sensors are becoming an increasingly common component of AVs’ sensor packages to provide redundancy and enhanced capabilities in conditions that are adverse for other types of sensors. For example, while RGB cameras are sensitive to lighting conditions and LiDAR performance is degraded in inclement weather such as rain, IR sensors are unaffected by lighting conditions and can contribute additional meaningful information in inclement weather. The US Army Corps of Engineers, Engineer Research and Development Center (ERDC) has developed the ERDC Computational Test Bed (CTB) to provide a suite of tools that can be used to support virtual development and testing of AVs. The CTB includes physics-based vehicle-terrain interaction, sensor and environment modeling, geo-environmental thermal modeling, software-inthe- loop capabilities, and virtual environment generation. Thermal modeling capabilities within the CTB utilize decades of near-surface phenomenology and autonomy research. Recent additions have been made to support large-domains commonly required for autonomous vehicle operations. These additions provide high-fidelity, physics-based thermal transfer and IR sensor models for creating high-quality synthetic imagery simulating IR sensors mounted on AVs. Highly parallelized thermal and IR sensor models for large-domain AV operations will be presented in this paper.
Statemental algebra and truth calculus for unmanned systemsChen, Xinjia
doi: 10.1117/12.2663294pmid: N/A
One crucial capability of unmanned systems is their ability to make decisions and inferences like humans. In this paper, we develop a novel logical system that imitates the way humans engage in reasoning with statements possessing varying degrees of ambiguity and unpredictability. Our proposed logical system is constructed using an axiomatic approach with self-evident rules, which allows us to define statemental operations and logical equivalence without the need for a concept of truth valuation. Our logical system includes both statemental algebra and truth calculus, which are designed to manipulate statements and assess their credibility. We believe that our proposed logical system has the potential to enhance the intelligence of unmanned systems.
Path-planning and obstacle avoidance algorithms for UAVs: a systematic literature reviewAlmazrouei, Khawla Saif; Bou Nassif, Ali; AlShabi, Mohammad
doi: 10.1117/12.2664056pmid: N/A
Recent interest in unmanned aerial vehicles (UAVs) has grown due to the wide range of possible civilian uses for these aircraft. However, present robot navigation technologies still need to be improved in various situations. Researchers are particularly interested in the 'Sense and Avoid' capacity as a critical issue. UAVs operating in civilian areas must have this functionality to do so safely. Numerous path planning and navigation algorithms have been developed for autonomous decision-making and control of UAVs. These path-planning algorithms are divided into either heuristic and non-heuristic or accurate methods. Both existing UAV route planning algorithms for the first and second techniques will be thoroughly compared in this work. Each algorithm is put through its paces in three diverse obstacle scenarios. Each method has been evaluated under various global and local obstacle information availability conditions while comparing the computational time and solution optimality.
Deep neural networks for detecting anomalies in unintended radiated emissionsWitham, Kenneth L.; Robb, Connor; Tahmoush, Dave
doi: 10.1117/12.2666271pmid: N/A
Unintended Radiated Emissions (URE) are emitted by all electrical devices and can be analyzed to determine changes in state on the emitting device. This paper aims to analyze UREs from a target device and classify if the device has changed operating states given a new measurement of the UREs. The UREs for a Raspberry Pi in different operating states are collected and analyzed. We detect if the target device changes operating states using a recurrent neural network to predict the URE spectrum power given multiple previous URE measurements. The predicted emissions are then compared to the measured emissions and a deep neural network classifies the measured emissions as a state change. Multiple other model types are compared including statistical classifiers and more complex machine learning models and our proposed model is found to perform the best in our dataset. We achieved an accuracy and F1-score of 90% in our real-world dataset.
An air-to-air unmanned aerial vehicle interceptor using machine-learning methods for detection and tracking of a target droneCheng, David; Nicol, Grant
doi: 10.1117/12.2663622pmid: N/A
The proliferation of small or micro unmanned aerial vehicles (UAV) gives rise to a potential threat for both public and military security. The small footprint and unpredictable dynamics of drones make detection and tracking difficult. Traditional methods of defence and protection may be ineffective against this new danger. This paper presents the work on developing DroneSwatter, a counter unmanned aerial system developed to track, follow, and take down a drone threat (Target Drone) using an agile, low-cost drone interceptor (Hunter Drone). The DroneSwatter project aims to apply machine learning techniques for counter-drone scenarios. Detection tasks are performed using deep learning detection algorithms. Simulation is used to build a tracking control model via proportional-derivative (PD) and machine learning algorithms. Optical pursuit based on images collected from the onboard camera of a Hunter Drone is implemented to track a Target Drone. Field experiments were conducted to test the feasibility and functionality of the current software and hardware methods for the DroneSwatter system. A benchmark was established by flying a target drone in designed patterns and the performance of the DroneSwatter tracking system was evaluated based on what speeds the Hunter Drone could follow the Target Drone in the field testing.
Initial investigation of UAV swarm behaviors in a search-and-rescue scenario using reinforcement learningCarley, Samantha S.; Price, Stanton R.; Hadia, Xian Mae D.; Price, Steven R.; Butler, Samantha J.
doi: 10.1117/12.2663629pmid: N/A
Recent years have seen the emergence of novel UAV swarm methodologies being developed for numerous applications within the Department of Defense. Such applications include, but are not limited to, search and rescue missions, intelligence, surveillance, and reconnaissance activities, and rapid disaster relief assessment. Herein, this article investigates an initial implementation of learning UAV swarm behaviors using reinforcement learning (RL). Specifically, we present a study implementing a leader-follower UAV swarm using RL-learned behaviors in a search-and-rescue task. Experiments are performed through simulations on synthetic data, specifically using a cross-platform flight simulator with Unreal Engine virtual environment. Performance is assessed by measuring key objective metrics, such as time to complete the mission, redundant actions, stagnation time, and goal success. This article seeks to provide an increased understanding and assessment of current reinforcement learning strategies being developed for controlling (or at a minimum suggesting) UAV swarm behaviors.