Nowadays, there are few unmanned aerial vehicles (UAVs) capable of flying, walking and grasping. A drone with all these functionalities can significantly improve its performance in complex tasks such as monitoring and exploring different types of terrain, and rescue operations. This paper presents MorphoArms, a novel system that consists of a morphogenetic chassis and a hand gesture recognition teleoperation system. The mechanics, electronics, control architecture, and walking behavior of the morphogenetic chassis are described. This robot is capable of walking and grasping objects using four robotic limbs. Robotic limbs with four degrees-of-freedom are used as pedipulators when walking and as manipulators when performing actions in the environment. The robot control system is implemented using teleoperation, where commands are given by hand gestures. A motion capture system is used to track the user's hands and to recognize their gestures. The method of controlling the robot was experimentally tested in a study involving 10 users. The evaluation included three questionnaires (NASA TLX, SUS, and UEQ). The results showed that the proposed system was more user-friendly than 56% of the systems, and it was rated above average in terms of attractiveness, stimulation, and novelty.
This paper presents a novel approach in autonomous robot control, named LLM-BRAIn, that makes possible robot behavior generation, based on operator's commands. LLM-BRAIn is a transformer-based Large Language Model (LLM) fine-tuned from Stanford Alpaca 7B model to generate robot behavior tree (BT) from the text description. We train the LLM-BRAIn on 8,5k instruction-following demonstrations, generated in the style of self-instruct using text-davinchi-003. The developed model accurately builds complex robot behavior while remaining small enough to be run on the robot's onboard microcomputer. The model gives structural and logical correct BTs and can successfully manage instructions that were not presented in training set. The experiment did not reveal any significant subjective differences between BTs generated by LLM-BRAIn and those created by humans (on average, participants were able to correctly distinguish between LLM-BRAIn generated BTs and human-created BTs in only 4.53 out of 10 cases, indicating that their performance was close to random chance). The proposed approach potentially can be applied to mobile robotics, drone operation, robot manipulator systems and Industry 4.0.
During operation, aerial manipulation systems are affected by various disturbances. Among them is a gravitational torque caused by the weight of the robotic arm. Common propeller-based actuation is ineffective against such disturbances because of possible overheating and high power consumption. To overcome this issue, in this paper we propose a winchbased actuation for the crane-stationed cable-suspended aerial manipulator. Three winch-controlled suspension rigging cables produce a desired cable tension distribution to generate a wrench that reduces the effect of gravitational torque. In order to coordinate the robotic arm and the winch-based actuation, a model-based hierarchical whole-body controller is adapted. It resolves two tasks: keeping the robotic arm end-effector at the desired pose and shifting the system center of mass in the location with zero gravitational torque. The performance of the introduced actuation system as well as control strategy is validated through experimental studies.
Visual localization is a fundamental task for a wide range of applications in the field of robotics. Yet, it is still a complex problem with no universal solution, and the existing approaches are difficult to scale: most state-of-the-art solutions are unable to provide accurate localization without a significant amount of storage space. We propose a hierarchical, low-memory approach to localization based on keypoints with different descriptor lengths. It becomes possible with the use of the developed unsupervised neural network, which predicts a feature pyramid with different descriptor lengths for images. This structure allows applying coarse-to-fine paradigms for localization based on keypoint map, and varying the accuracy of localization by changing the type of the descriptors used in the pipeline. Our approach achieves comparable results in localization accuracy and a significant reduction in memory consumption (up to 16 times) among state-of-the-art methods.
In this paper, we propose a novel approach to wheeled robot navigation through an environment with movable obstacles. A robot exploits knowledge about different obstacle classes and selects the minimally invasive action to perform to clear the path. We trained a convolutional neural network (CNN), so the robot can classify an RGB-D image and decide whether to push a blocking object and which force to apply. After known objects are segmented, they are being projected to a cost-map, and a robot calculates an optimal path to the goal. If the blocking objects are allowed to be moved, a robot drives through them while pushing them away. We implemented our algorithm in ROS, and an extensive set of simulations showed that the robot successfully overcomes the blocked regions. Our approach allows a robot to successfully build a path through regions, where it would have stuck with traditional path-planning techniques.
The continuous monitoring by drone swarms remains a challenging problem due to the lack of power supply and the inability of drones to land on uneven surfaces. Heterogeneous swarms, including ground and aerial vehicles, can support longer inspections and carry a higher number of sensors on board. However, their capabilities are limited by the mobility of wheeled and legged robots in a cluttered environment. In this paper, we propose a novel concept for autonomous inspection that we call SwarmGear. SwarmGear utilizes a heterogeneous swarm that investigates the environment in a leader-follower formation. The leader drone is able to land on rough terrain and traverse it by four compliant robotic legs, possessing both the functionalities of an aerial and mobile robot. To preserve the formation of the swarm during its motion, virtual impedance links were developed between the leader and the follower drones. We evaluated experimentally the accuracy of the hybrid leader drone's ground locomotion. By changing the step parameters, the optimal step configuration was found. Two types of gaits were evaluated. The experiments revealed low crosstrack error (mean of 2 cm and max of 4.8 cm) and the ability of the leader drone to move with a 190 mm step length and a 3 degree standard yaw deviation. Four types of drone formations were considered. The best formation was used for experiments with SwarmGear, and it showed low overall crosstrack error for the swarm (mean 7.9 cm for the type 1 gait and 5.1 cm for the type 2 gait). The proposed system can potentially improve the performance of autonomous swarms in cluttered and unstructured environments by allowing all agents of the swarm to switch between aerial and ground formations to overcome various obstacles and perform missions over a large area.
This work presents an RL-based agent for outpatient hysteroscopy training. Hysteroscopy is a gynecological procedure for examination of the uterine cavity. Recent advancements enabled performing this type of intervention in the outpatient setup without anaesthesia. While being beneficial to the patient, this approach introduces new challenges for clinicians, who should take additional measures to maintain the level of patient comfort and prevent tissue damage. Our prior work has presented a platform for hysteroscopic training with the focus on the passage of the cervical canal. With this work, we aim to extend the functionality of the platform by designing a subsystem that autonomously performs the task of the passage of the cervical canal. This feature can later be used as a virtual instructor to provide educational cues for trainees and assess their performance. The developed algorithm is based on the soft actor critic approach to smooth the learning curve of the agent and ensure uniform exploration of the workspace. The designed algorithm was tested against the performance of five clinicians. Overall, the algorithm demonstrated high efficiency and reliability, succeeding in 98% of trials and outperforming the expert group in three out of four measured metrics.
Visual localization is an essential modern technology for robotics and computer vision. Popular approaches for solving this task are image-based methods. Nowadays, these methods have low accuracy and a long training time. The reasons are the lack of rigid-body and projective geometry awareness, landmark symmetry, and homogeneous error assumption. We propose a heterogeneous loss function based on concentrated Gaussian distribution with the Lie group to overcome these difficulties. Following our experiment, the proposed method allows us to speed up the training process significantly (from 300 to 10 epochs) with acceptable error values.
This paper presents a novel haptic device DeltaFinger designed to deliver the force of interaction with virtual objects by guiding user's finger with wearable delta mechanism. The developed interface is capable to deliver 3D force vector to the fingertip of the index finger of the user, allowing complex rendering of virtual reality (VR) environment. The developed device is able to produce the kinesthetic feedback up to 1.8 N in vertical projection and 0.9 N in horizontal projection without restricting the motion freedom of of the remaining fingers. The experimental results showed a sufficient precision in perception of force vector with DeltaFinger (mean force vector error of 0.6 rad). The proposed device potentially can be applied to VR communications, medicine, and navigation of the people with vision problems.
We propose a novel concept of augmented reality (AR) human-drone interaction driven by RL-based swarm behavior to achieve intuitive and immersive control of a swarm formation of unmanned aerial vehicles. The DroneARchery system developed by us allows the user to quickly deploy a swarm of drones, generating flight paths simulating archery. The haptic interface LinkGlide delivers a tactile stimulus of the bowstring tension to the forearm to increase the precision of aiming. The swarm of released drones dynamically avoids collisions between each other, the drone following the user, and external obstacles with behavior control based on deep reinforcement learning. The developed concept was tested in the scenario with a human, where the user shoots from a virtual bow with a real drone to hit the target. The human operator observes the ballistic trajectory of the drone in an AR and achieves a realistic and highly recognizable experience of the bowstring tension through the haptic display. The experimental results revealed that the system improves trajectory prediction accuracy by 63.3% through applying AR technology and conveying haptic feedback of pulling force. DroneARchery users highlighted the naturalness (4.3 out of 5 point Likert scale) and increased confidence (4.7 out of 5) when controlling the drone. We have designed the tactile patterns to present four sliding distances (tension) and three applied force levels (stiffness) of the haptic display. Users demonstrated the ability to distinguish tactile patterns produced by the haptic display representing varying bowstring tension(average recognition rate is of 72.8%) and stiffness (average recognition rate is of 94.2%). The novelty of the research is the development of an AR-based approach for drone control that does not require special skills and training from the operator.