ICRA 2020 Papers

SwarmRail represents a novel solution to overhead manipulation from a mobile unit that drives in an aboveground rail-structure. The concept is based on the combination of omnidirectional mobile platform and L-shaped rail profiles that form a through-going central gap. This gap makes possible mounting a robotic manipulator arm overhead at the underside of the mobile platform. Compared to existing solutions, SwarmRail enables continuous overhead manipulation while traversing rail crossings. It also can be operated in a robot swarm, as it allows for concurrent operation of a group of mobile SwarmRail units inside a single rail network. Experiments on a first functional demonstrator confirm the functional capability of the concept. Potential fields of applications reach from industry over logistics to vertical farming.

We present an approach to the distributed storage of data across a swarm of mobile robots that forms a shared global memory. We assume that external storage infrastructure is absent, and that each robot is capable of devoting a quota of memory and bandwidth to distributed storage. Our approach is motivated by the insight that in many applications data is collected at the periphery of a swarm topology, but the periphery also happens to be the most dangerous location for storing data, especially in exploration missions. Our approach is designed to promote data storage in the locations in the swarm that best suit a specific feature of interest in the data, while accounting for the constantly changing topology due to individual motion. We analyze two possible features of interest: the data type and the data item position in the environment. We assess the performance of our approach in a large set of simulated experiments. The evaluation shows that our approach is capable of storing quantities of data that exceed the memory of individual robots, while maintaining near-perfect data retention in high-load conditions.

Precision cutting of soft-tissue remains a challenging problem in robotics, due to the complex and unpredictable mechanical behaviour of tissue under manipulation. Here, we consider the challenge of cutting along the boundary between two soft mediums, a problem that is made extremely difficult due to visibility constraints, which means that the precise location of the cutting trajectory is typically unknown. This paper introduces a novel strategy to address this task, using a binary medium classifier trained using joint torque measurements, and a closed loop control law that relies on an error signal compactly encoded in the decision boundary of the classifier. We illustrate this on a grapefruit cutting task, successfully modulating a nominal trajectory t using dynamic movement primitives to follow the boundary between grapefruit pulp and peel using torque based medium classification. Results show that this control strategy is successful in 72 % of attempts in contrast to control using a nominal trajectory, which only succeeds in 50 % of attempts.

Supervisory Control Theory (SCT) provides a formal framework for controlling discrete event systems. It has recently been used to generate correct-by-construction controllers for swarm robotics systems. Current SCT frameworks are limited, as they support only (private) events that are observable within the same robot. In this paper, we propose an extended SCT framework that incorporates (public) events that are shared among robots. The extended framework allows to model formally the interactions among the robots. It is evaluated using a case study, where a group of mobile robots need to synchronise their movements in space and time-a requirement that is specified at the formal level. We validate our approach through experiments with groups of e-puck robots.

Depth acquisition, based on active illumination, is essential for autonomous and robotic navigation. LiDARs (Light Detection And Ranging) with mechanical, fixed, sampling templates are commonly used in today's autonomous vehicles. An emerging technology, based on solid-state depth sensors, with no mechanical parts, allows fast and adaptive scans. In this paper, we propose an adaptive, image-driven, fast, sampling and reconstruction strategy. First, we formulate a piece-wise planar depth model and estimate its validity for indoor and outdoor scenes. Our model and experiments predict that, in the optimal case, adaptive sampling strategies with about 20-60 piece-wise planar structures can approximate well a depth map. This translates to requiring a single depth sample for every 1200 RGB samples (less than 0.1%), providing strong motivation to investigate an adaptive framework. Second, we introduce SPS (Super-Pixel Sampler), a simple, generic, sampling and reconstruction algorithm, based on super-pixels. Our sampling improves grid and random sampling, consistently, for a wide variety of reconstruction methods. Third, we propose an extremely simple and fast reconstruction for our sampler. It achieves state-of-the-art results, compared to complex image- guided depth completion algorithms, reducing the required sampling rate by a factor of 3-4. A single-pixel prototype sampler built in our lab illustrates the concept.

Modeling how a robot interacts with the environment around it is an important prerequisite for designing control and planning algorithms. In fact, the performance of controllers and planners is highly dependent on the quality of the model. One popular approach is to learn data driven models in order to compensate for inaccurate physical measurements and to adapt to systems that evolve over time. In this paper, we investigate a method to regularize model learning techniques to provide better error characteristics for traditional control and planning algorithms. This work proposes learning "Sufficiently Accurate" models of dynamics using a primal-dual method that can explicitly enforce constraints on the error in pre-defined parts of the state-space. The result of this method is that the error characteristics of the learned model is more predictable and can be better utilized by planning and control algorithms. The characteristics of Sufficiently Accurate models are analyzed through experiments on a simulated ball paddle system.

For safe human-robot interaction, various type of flexible manipulators have been developed. Especially series elastic actuator (SEA) based manipulators have been getting huge attention since the elastic element of SEA prevents people from injury when undesirable collision happens. Moreover, it improves system durability by absorbing impact force, which could damage actuators. However, the elastic element inside SEA manipulator causes low system bandwidth which limits the speed performance of conventional impedance control approaches. To alleviate the low bandwidth issue of impedance controlled SEA while guaranteeing system stability, we implement Time Domain Passivity Approach (TDPA) and Successive Stiffness Increment (SSI) approach, which was invented in haptic and teleoperation domain. Impedance controlled SEA is reformulated as a two-port electrical circuit network for implementing TDPA. In addition, a pair of input and output power conjugate variable, dominating the system passivity is identified for implementing SSI approach. Experimental results showed that TDPA and SSI approach can render the stiffness of the impedance controller, which decides the bandwidth, upto 350 kN/m without any stability issue, while normal impedance controller only render upto 120 kN/m. Although both of the approaches significantly increased the bandwidth of the impedance controlled SEA, TDPA slightly outperformed in stability, and SSI outperformed in tracking.

In this paper, we study the state-constrained consensus problem and introduce a new family of distributed algorithms based on subspace projection methods which are simple to implement and which preserve, under some suitable conditions, the consensus value of the original discrete-time agreement protocol. The proposed theory is supported by extensive numerical experiments for the constrained 2D rendezvous of single-integrator robots.

Social Navigation methods attempt to integrate knowledge from Human Sciences fields such as the notion of Proxemics into mobile robot navigation. They are often evaluated in simulations, or lab conditions with informed participants, and studies of the impact of the robot behavior on humans are rare. Humans communicate and interact through many vectors, among which are motion and positioning, which can be related to social hierarchy and the socio-physical context. If a robot is to be deployed among humans, the methods it uses should be designed with this in mind. This work acts as the first step in an ongoing project in which we explore how to design navigation methods for mobile robots destined to be deployed among humans. We aim to consider navigation as more than just a functionality of the robot, and to study the impact of robot motion on humans. In this paper, we focus on the person-following task. We selected a state of the art person-following method as the basis for our method, which we modified and extended in order for it to be more general and adaptable. We conducted pilot experiments using this method on a real mobile robot in ecological contexts. We used results from the experiments to study the Human-Robot Interaction as a whole by analysing both the person-following method and the human behavior. Our preliminary results show that the way in which the robot followed a person had an impact on the interaction that emerged between them.

This paper presents the architecture of a high- density parking solution based on car-like robots specifically designed to move cars. The main difficulty is to park the vehicles close to one another which implies hard constraints on the robot motion and localization. In particular, this paper focuses on navigation in narrow lanes. We propose a Lyapunov- based control strategy that has been derived after expressing the problem in a Configuration Space formulation. The current solution has been implemented and tested on Stanley Robotics' robots and has been running in production for several months. Thanks to the Configuration Space formulation, we are able to guarantee the obstacles’ integrity. Moreover, a method for calibrating the GPS orientation with a high-precision is derived from the present control strategy.

Training data is the key ingredient for deep learning approaches, but difficult to obtain for the specialized domains often encountered in robotics. We describe a synthesis pipeline capable of producing training data for cluttered scene perception tasks such as semantic segmentation, object detection, and correspondence or pose estimation. Our approach arranges object meshes in physically realistic, dense scenes using physics simulation. The arranged scenes are rendered using high-quality rasterization with randomized appearance and material parameters. Noise and other transformations introduced by the camera sensors are simulated. Our pipeline can be run online during training of a deep neural network, yielding applications in life-long learning and in iterative render-and-compare approaches. We demonstrate the usability by learning semantic segmentation on the challenging YCB-Video dataset without actually using any training frames, where our method achieves performance comparable to a conventionally trained model. Additionally, we show successful application in a real-world regrasping system.

In this paper, the optimization of the anchor points of a cable driven parallel robot (CDPR) for 3D printing is proposed in order to maximize the rigidity. Indeed, in the context of 3D printing, robot stiffness should guarantee a high level of tool path following accuracy. The optimized platform showed a rigidity improvement in simulation, but also experimentally with a first study of vibration modes. In the same time, this study illustrates the influence of preload in cables on the platform rigidity.

Stereo-vision devices have rigorous requirements for extrinsic parameter calibration. In Stereo Visual Inertial Odometry (VIO), inaccuracy in or changes to camera extrinsic parameters may lead to serious degradation in estimation performance. In this manuscript, we propose an online calibration method for stereo VIO extrinsic parameters correction. In particular, we focus on Multi-State Constraint Kalman Filter (MSCKF [1]) framework to implement our method. The key component is to formulate stereo extrinsic parameters as part of the state variables and model the Jacobian of feature reprojection error with respect to stereo extrinsic parameters as sub-block of update Jacobian. Therefore we can estimate stereo extrinsic parameters simultaneously with inertial measurement unit (IMU) states and camera poses. Experiments on EuRoC dataset and real-world outdoor dataset demonstrate that the proposed algorithm produce higher positioning accuracy than the original S-MSCKF [2], and the noise of camera extrinsic parameters are self-corrected within the system.

Artificial microswimmers are prospective robotic agents especially in biomedical applications. A rotating magnetic field can actuate a magnetized swimmer with a helical tail and enable propulsion. Such swimmers exhibit several modes of instability. Inside conduits, for example, hydrodynamic interactions with the boundaries lead to helical paths for pusher-mode swimmers; in this mode the helical tail pushes a rotating magnetic head. State-of-the-art in controlled navigation of micro-swimmers is based on aligning the swimmer orientation according to a reference path, thereby requiring both swimmer orientation and position to be known. Object-orientation is hard to track especially in in vivo scenarios which render orientation-based methods practically unfeasible. Here, we show that the kinematics for a confined swimmer can be linearized by assuming a low wobbling angle. This allows for a control law solely based on the swimmer position. The approach is demonstrated through experiments and two different numerical models: the first is based on the resistive force theory for a swimmer inside a swirling flow represented by a forced vortex and the second is a computational fluid dynamics model, which solves Stokes equations for a swimmer inside a circular channel. Helical pusher-mode trajectories are suppressed significantly for the straight path following problem. The error in real-life experiments remains comparable to those in the state-of-the-art methods.

This paper investigates a burrowing robot that can maneuver and steer while being submerged in a granular medium. The robot locomotes using an internal vibro-impact mechanism and steers using a rotating bevel-tip head. We formulate and investigate a non-holonomic model for the steering mechanism and a hybrid dynamics model for the thrusting mechanism. We perform a numerical analysis of the dynamics of the robot's thrusting mechanism using a simplified, orientation and depth dependent model for the drag forces acting on the robot. We first show, in simulation, that by carefully tuning various control input parameters, the thrusting mechanism can drive the robot both forward and backward. We present several experiments designed to evaluate and verify the simulative results using a proof-of-concept robot. We show that different input amplitudes indeed affect the direction of motion, as suggested by the simulation. We further demonstrate the ability of the robot to perform a simple S-shaped trajectory. These experiments demonstrate the feasibility of the robot's design and fidelity of the model.

Advances in legged robotics are strongly rooted in animal observations. A clear illustration of this claim is the generalization of Central Pattern Generators (CPG), first identified in the cat spinal cord, to generate cyclic motion in robotic locomotion. Despite a global endorsement of this model, physiological and functional experiments in mammals have also indicated the presence of descending signals from the cerebellum, and reflex feedback from the lower limb sensory cells, that closely interact with CPGs. To this day, these interactions are not fully understood. In some studies, it was demonstrated that pure reflex-based locomotion in the absence of oscillatory signals could be achieved in realistic musculoskeletal simulation models or small compliant quadruped robots. At the same time, biological evidence has attested the functional role of the cerebellum for predictive control of balance and stance within mammals. In this paper, we promote both approaches and successfully apply reflex-based dynamic locomotion, coupled with a balance and gravity compensation mechanism, on the state-of-art HyQ robot. We discuss the importance of this stability module to ensure a correct foot lift-off and maintain a reliable gait. The robotic platform is further used to test two different architectural hypotheses inspired by the cerebellum. An analysis of experimental results demonstrates that the most biologically plausible alternative also leads to better results for robust locomotion.

The flexible under-actuated musculoskeletal hand is superior in its adaptability and impact resistance. On the other hand, since the relationship between sensors and actuators cannot be uniquely determined, almost all its controls are based on feedforward controls. When grasping and using a tool, the contact state of the hand gradually changes due to the inertia of the tool or impact of action, and the initial contact state is hardly kept. In this study, we propose a system that trains the predictive network of sensor state transition using the actual robot sensor information, and keeps the initial contact state by a feedback control using the network. We conduct experiments of hammer hitting, vacuuming, and brooming, and verify the effectiveness of this study.

Stable flight control under ceilings is difficult for multirotor Unmanned Aerial Vehicles (UAVs). The wake interaction between rotors and upper walls, called the "ceiling effect", causes an increase of rotor thrust. As a result of the thrust increase, multi-rotors are drawn upward abruptly and collide with ceilings. In previous work, several thrust models of the ceiling effect have been proposed for stable flight under ceilings, assuming that the airflow around rotors is in steady states. However, the airflow around rotors in vertical flight is not in steady states and each thrust model in previous work is skillfully determined based on large amounts of precise experimental data. In this paper, we introduce an aerodynamics-based thrust model and a stable control method under ceilings. This model is derived from the momentum theory and the relationship between vertical climbing/descending rates of rotors and an induced velocity. To confirm our proposed model, we collect thrust data at various vertical rates in flight. In addition, we use only onboard sensors to estimate selfstate for structural inspections. Consequently, we reveal that the proposed model is consistent with the experimental results. Based on an aerodynamic model, we need not collect large amounts of precise experimental data to realize stable flight. Furthermore, the vertical flight tests under ceilings demonstrate that our in-unsteady-state-model-based controller outperforms the conventional steady-state ones.

Biped walking involves a series of transitions between single support (SS) and double support (DS) contact configurations that can include both balanced and unbalanced states. The new concept of steppability is introduced to partition the set of unbalanced states into steppable states and falling (unsteppable) states based on the ability of a biped system to respond to forward velocity perturbations by stepping. In this work, a complete system-specific analysis of the stepping process including full-order nonlinear system dynamics is presented for the DARwIn-OP humanoid robot and a human subject in the sagittal plane with respect to both balance stability and steppability. The balance stability and steppability of each system are analyzed by numerical construction of its balance stability boundaries (BSB) for the initial SS and final DS contact configuration and the steppable unbalanced state boundary (SUB). These results are presented with center of mass (COM) trajectories obtained from walking experiments to benchmark robot controller performance and analyze the variation of balance stability and steppability with COM and swing foot position along the progression of a step cycle. For each system, DS BSBs were obtained with both constrained and unconstrained arms in order to demonstrate the ability of this approach to incorporate the effects of angular momentum and system-specific characteristics such as actuation torque, velocity, and angle limits.