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.
Space debris in the 1-10 cm diameter regime presents a particular hazard to future operations in low Earth orbit (LEO). These objects are too small to be reliably detected by terrestrial radar or telescopes, but too large for their impacts to be mitigated with Whipple shielding. This paper addresses this hazard by presenting a methodology for space-based detection and active debris removal (ADR) via novel intelligent collaborative sensing and efficient decentralized data fusion. In simulation, we demonstrate that a small heterogeneous constellation of co-orbiting satellites using attentional neural networks can autonomously share information and maneuver in a cooperative game to de-orbit space debris (e.g. with a mechanical arm or pulsed laser) while minimizing total energy consumption and collision risk in the constellation. Particular focus is given to the robustness and effectiveness of the developed collaborative sensing system with respect to sensor uncertainties and transient changes in observability between nodes.
Rise of unmanned vehicles and autonomous robots has been accompanied by study of path planning, navigation, and decision-making algorithms. Current state-of-the-art employs deep neural nets to extract the required features. This technology, though successful in most cases, fails where the training has not been done for unseen conditions. For such unlabeled training data, transfer learning approaches have been proposed. A major drawback of using transfer learning approaches is that the actions and/or state spaces are reactive only to present circumstances. A truly intelligent autonomous operation has to consider a subordinate-to-a-human approach for its mission risks that vary with topography, path planning as well as mission goals. To address these complex combinatorial problems, DARPA has initiative a novel Explainable AI (XAI) technology in the past few years. In XAI, machine learning is paired with human intervention to make decisions by generating textual explanations of all the available relevant information / decision. In this paper, we propose to use available information along with human intelligence in a feedback loop for helping the unlabeled data to be trained and generate cost functions which were previously not programmed. We study this context/situation awareness to generate list of decision available from explanations on combinatorial tasks. Moreover, we employ this approach to a quadruped robot to learn its environment and the AI model starts in its infancy to mimic human cognitive architecture. We show that the learning process can be improved in a way that suits a particular mission in mind.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.