We propose a novel real-time excavation trajectory modulation framework on a slope for an autonomous excavator with a low-level digital kinematic control as common for hydraulic industrial excavators. Excavation on a slope is challenging because of a higher risk of slips and rollovers. To deal with this, we propose a real-time excavation trajectory modulation framework based on slope tangential/normal force ratio mu and zero moment point xi. The slip and rollover prevention conditions are incorporated in a single linear inequality using the same fractional structure in mu and xi with the common denominator. However, due to the adoption of the low-level digital kinematic control, this prevention requires the prediction of the excavation force at the next timestamp, and, for this, we develop a data-driven excavation force difference prediction model utilizing a deep learning architecture, Transformer. The remaining error of this prediction is then addressed by using the technique of robust optimization with box uncertainty of the developed excavation force difference model. Our proposed framework is validated experimentally with our customized scaled-down excavator.

Simulation with a reasonable physical model is important to develop control algorithms for robots quickly, accurately, and safely without damaging the associated physical systems in various environments. However, it is difficult to choose the suitable tool for simulating a specific project. To help users in selecting the best tool when simulating a given project, we compare the performance of the four widely used physics engines, namely, ODE, Bullet, Vortex, and MoJoco, for various simple and complex industrial scenarios. We first summarize the technical algorithms implemented in each physics engine. We also designed four simulation scenarios ranging from simple scenarios for which analytic solution exists to complex industrial scenarios to compare the performance of each physics engine. We then present the simulation results in the default settings of all the physics engines, and analyze the behavior and contact force of the simulated objects.

We present a two-stage framework that integrates a learning-based estimator and a controller, designed to address contact-intensive tasks. The estimator leverages a Bayesian particle filter with a mixture density network (MDN) structure, effectively handling multi-modal issues arising from contact information. The controller combines a self-supervised and reinforcement learning (RL) approach, strategically dividing the low-level admittance controller's parameters into labelable and non-labelable categories, which are then trained accordingly. To further enhance accuracy and generalization performance, a transformer model is incorporated into the self-supervised learning component. The proposed framework is evaluated on the bolting task using an accurate real-time simulator and successfully transferred to an experimental environment. More visualization results are available on our project website: https://sites.google.com/view/2stagecitt

There is a growing interest in learning a velocity command tracking controller of quadruped robot using reinforcement learning due to its robustness and scalability. However, a single policy, trained end-to-end, usually shows a single gait regardless of the command velocity. This could be a suboptimal solution considering the existence of optimal gait according to the velocity for quadruped animals. In this work, we propose a hierarchical controller for quadruped robot that could generate multiple gaits (i.e. pace, trot, bound) while tracking velocity command. Our controller is composed of two policies, each working as a central pattern generator and local feedback controller, and trained with hierarchical reinforcement learning. Experiment results show 1) the existence of optimal gait for specific velocity range 2) the efficiency of our hierarchical controller compared to a controller composed of a single policy, which usually shows a single gait. Codes are publicly available.

We propose a novel excavation (i.e., digging) trajectory planning framework for industrial autonomous robotic excavators, which emulates the strategies of human expert operators to optimize the excavation of (complex/unmodellable) soils while also upholding robustness and safety in practice. First, we encode the trajectory with dynamic movement primitives (DMP), which is known to robustly preserve qualitative shape of the trajectory and attraction to (variable) end-points (i.e., start-points of swing/dumping), while also being data-efficient due to its structure, thus, suitable for our purpose, where expert data collection is expensive. We further shape this DMPbased trajectory to be expert-emulating, by learning the shaping force of the DMP-dynamics from the real expert excavation data via a neural network (i.e., MLP (multi-layer perceptron)). To cope with (possibly dangerous) underground uncertainties (e.g., pipes, rocks), we also real-time modulate the expert-emulating (nominal) trajectory to prevent excessive build-up of excavation force by using the feedback of its online estimation. The proposed framework is then validated/demonstrated by using an industrial-scale autonomous robotic excavator, with the associated data also presented here.

Haptic interface technologies for virtual reality applications have been developed to increase the reality and manipulability of a virtual object by creating a diverse tactile sensation. Most evaluation of the haptic technologies, however, have been limited to the haptic perception of the tactile stimuli via static virtual objects. Noting this, we investigated the effect of lateral cutaneous feedback, along with kinesthetic feedback on the perception of virtual object weight during manipulation. We modeled the physical interaction between a participant’s finger avatars and virtual objects. The haptic stimuli were rendered with custom-built haptic feedback systems that can provide kinesthetic and lateral cutaneous feedback to the participant. We conducted two virtual object manipulation experiments, 1. a virtual object manipulation with one finger, and 2. the pull-out and lift-up of a virtual object grasped with a precision grip. The results of Experiment 1 indicate that the participants felt the virtual object rendered with lateral cutaneous feedback significantly heavier than with only kinesthetic feedback ( p < 0.05 for m ref = 100 and 200 g). Similarly, the participants of Experiment 2 felt the virtual objects significantly heavier when lateral cutaneous feedback was available ( p < 0.05 for m ref = 100, 200, and 300 g). Therefore, the additional lateral cutaneous feedback to the force feedback led the participants to feel the virtual object heavier than without the cutaneous feedback. The results also indicate that the contact force applied to a virtual object during manipulation can be a function of the perceived object weight ( p = 0.005 for Experiment 1 and p = 0.02 for Experiment 2).

This paper presents a new actuator mechanism that can create a planar two-DOF impact and vibrotactile stimuli with wide frequency bandwidth. The proposed device utilizes one permanent magnet and two different sets of solenoids, direction controlling solenoids and a central solenoid. The direction controlling solenoids are two sets of facing solenoid pair, which attract the permanent magnet from a neutral position toward the housing, to create an impact. The central solenoid increases the magnet's potential energy, resulting in a faster movement toward the outside of the neutral position and thus increasing the amount of impact. The central solenoid can be utilized to create an attractive force to move the magnet back to the neutral position after the impact. When creating an impact in an arbitrary orientation, two sets of solenoid pair are used to decide the initial force direction, which is parallel to the magnet's trajectory until an impact. The actuator can be used as a tactor by controlling the impact cycle.

This paper presents a glove type haptic interface that can provide cutaneous feedback to the dorsum of a user's hand. The interface consists of flex sensors, a position tracker and skin stretch modules which provides cutaneous feedback to the hand. Global position of the wrist is tracked with the position tracker and the finger posture is estimated with the flex sensors. Two skin-stretch modules installed on the dorsum of the hand behind the MCP joint of the thumb, and the middle of the index and middle fingers. Whenever there is a contact between a fingertip and a virtual object, the skin-stretch module provides cutaneous feedback by rotating a contact element toward the wrist. The skin-stretch emulates the strain of the skin occurring when a finger is pushed away by touching a real object.

When one manipulates a large or bulky object, s/he utilizes tactile information at both fingers and the palm. Our goal is to efficiently convey contact information to a user's hand during interaction with a virtual object. We propose a haptic system that can provide haptic feedback to thumb/middle finger/index finger and on a palm. Our interface design utilizes a novel compact mechanism to provide haptic information to the palm. Also, we propose a haptic rendering strategy to calculate haptic feedback continuously. We demonstrate that cutaneous feedback on the palm improves the haptic perception of a large virtual object compared to when there is only kinesthetic feedback to the fingers.

Tactile information in a palm is a necessary component in manipulating and perceiving large or heavy objects. Noting this, we investigate human sensitivity to tactile haptic feedback in a palm for an improved user interface design. To provide distributed tactile pattern, we propose an ungrounded haptic interface, which can stimulate multiple locations in a palm, independently. Two experiments were conducted to evaluate human sensitivity to distributed tactile patterns. The first experiment tested participants' sensitivity to tactile patterns by sub-sections in a palm, and a significant effect of the sub-section on the sensitivity was observed. In the second experiment, participants identified pressure distribution patterns in the palm collected from real-life objects with the percent correct of 71.4 % and IT (information transfer) was 1.58 bits.