IROS 2019 Papers

Electric aerial propulsion systems are commonly used for many small-scale unmanned aerial vehicles (UAVs), providing a light and powerful method of generating thrust. In the emerging area of aerial-aquatic vehicles, most existing prototypes rely on such systems to propel themselves in both air and water. As the density of water is three orders of magnitude larger than that of air, a spinning aerodynamic body in the medium will experience significantly higher torque at the same speed. This results in aerial propulsion systems to be heavily mismatched underwater, as the required torque is higher than the drive torque that a typical aerial motor can provide. Here, an in-depth investigation of such off-design operation is conducted. Based on numerical simulation, we identify the feasible operating range of such systems and present an evaluation framework that identifies a motor-propeller combination from a component database that maximises underwater performance while ensuring aerial thrust requirements are met.

We discuss motion planning in the configuration spaces of robots containing continuum elements. The configuration space structure of extensible continuum sections is first analyzed, with practical constraints unique to continuum elements identified. The results are applied to generate the configuration space of a hybrid continuum lamp/mobile base robot. A conventional motion planning RRT/A* approach is subsequently applied for the robot in an aging in place application scenario.

In this paper, 3D humanoid walking is decoupled into periodic and transitional motion, each of which is decoupled into planar walking in the sagittal and lateral plane. Reduced order models (ROMs), i.e. actuated Spring-loaded Inverted Pendulum (aSLIP) models and Hybrid-Linear Inverted Pendulum (H-LIP) models, are utilized for motion generation on the desired center of mass (COM) dynamics for each type of planar motion. The periodic motion is planned via point foot (underactuated) ROMs for dynamic motion with minimum ankle actuation, while the transitional motion is planned via foot-actuated ROMs for fast and smooth transition. Composition of the planar COM dynamics yields the desired COM dynamics in 3D, which is embedded on the humanoid via control Lyapunov function based Quadratic programs (CLF-QPs). Additionally, the ground reaction force profiles of the aSLIP walking are used as desired references for ground contact forces in the CLF-QPs for smooth domain transitions. The proposed framework is realized on a lower-limb exoskeleton in simulation wherein different walking motions are achieved.

Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into a bird's eye view, also known as plan view, for end-to-end control. We detect vehicles and pedestrians in the first person view and project them into an overhead plan view. This representation provides an abstraction of the environment from which a deep network can easily deduce the positions and directions of entities. Additionally, the plan view enables us to leverage advances in 3D object detection in conjunction with deep policy learning. We evaluate our monocular plan view network on the photo-realistic Grand Theft Auto V simulator. A network using both a plan view and front view causes less than half as many collisions as previous detection-based methods and an order of magnitude fewer collisions than pure pixel-based policies.

In many robotic applications, especially for the autonomous driving, understanding the semantic information and the geometric structure of surroundings are both essential. Semantic 3D maps, as a carrier of the environmental knowledge, are then intensively studied for their abilities and applications. However, it is still challenging to produce a dense outdoor semantic map from a monocular image stream. Motivated by this target, in this paper, we propose a method for large-scale 3D reconstruction from consecutive monocular images. First, with the correlation of underlying information between depth and semantic prediction, a novel multi-task Convolutional Neural Network (CNN) is designed for joint prediction. Given a single image, the network learns low-level information with a shared encoder and separately predicts with decoders containing additional Atrous Spatial Pyramid Pooling (ASPP) layers and the residual connection which merits disparities and semantic mutually. To overcome the inconsistency of monocular depth prediction for reconstruction, post-processing steps with the superpixelization and the effective 3D representation approach are obtained to give the final semantic map. Experiments are compared with other methods on both semantic labeling and depth prediction. We also qualitatively demonstrate the map reconstructed from large-scale, difficult monocular image sequences to prove the effectiveness and superiority.

In this work we propose an unsupervised training method that finetunes a single image depth estimation CNN towards a new environment. The network, which has been pretrained on stereo data, only requires monocular input for finetuning. Unlike other unsupervised methods, it produces depth estimations with absolute scale - a feature that is essential for most practical applications, yet has mostly been overlooked in the literature. First, we show how our method allows adapting a network trained on one dataset (Cityscapes) to another (KITTI). Next, by splitting KITTI in subsets, we show the sensitivity of pretrained models to a domain shift. We then demonstrate that, by finetuning the model using our method, it is possible to improve the performance on the target subset, without using stereo or any form of groundtruth depth and with preservation of the correct absolute scale.

Volume change has become a critical actuation method in robotics. However, the need for fluid flow or thermal processes to generate volume changes limits the durability, speed, and efficiency of these actuators. In this paper, we develop a new electromechanical actuator that volumetrically expands. By combining auxetic materials with a servo, we produce a simple isotropically expanding actuator that can be modularly composed. We discuss the symmetry considerations in selecting an appropriate auxetic framework for our actuator, eventually choosing a double-layered polyhedral auxetic design. Characterization shows that a single actuator can expand in radius to 119% of the original size and generate 90N of force, while maintaining a small package and a speedy expansion / contraction cycle. Finally, we demonstrate the modularity of our actuators by linking three actuators to create a vertical tube-crawling robot. The small package and fast cycle time of our system highlight how viable these electromechanical volumetric actuators can be as an important actuator modality.

Soft strain resistive sensors based on eutectic gallium-indium liquid metal can play an important role in proprioceptive sensing for soft robots. However, there are no available mathematical models to accurately estimate the strain as a function of the measured resistance. Furthermore, non-uniform strain in the microchannels has not been analysed yet. In this paper, we introduce a new model to estimate the strain or elongation in sub-millimetre scale, and analyse its accuracy through a customised testing set-up and procedure. The effect of strain rate on the measurement accuracy is also studied. We compare existing theoretical models with our experimental results, and discuss the differences between them. Moreover, we analyse the effect of strain rate on hysteresis caused by the viscoelastic behaviour and introduce a new model for it to be potentially used for future work. This paper demonstrates, among other things, that rational models could provide high accuracy in strain estimation, and might help to enhance proprioceptive sensing and state control of soft robots.

Existing works on control of tractor-trailers systems only consider the kinematics model without taking dynamics into account. Also, most of them treat the issue as a pure control theory problem whose solutions are difficult to implement. This paper presents a trajectory tracking control approach for a full-scale industrial tractor-trailers vehicle composed of a carlike tractor and arbitrary number of passive full trailers. To deal with dynamic effects of trailing units, a force sensor is innovatively installed at the connection between the tractor and the first trailer to measure the forces acting on the tractor. The tractor's dynamic model that explicitly accounts for the measured forces is derived. A tracking controller that compensates the pulling/pushing forces in real time and simultaneously drives the system onto desired trajectories is proposed. The propulsion map between throttle opening and the propulsion force is proposed to be modeled with a fifth-order polynomial. The parameters are estimated by fitting experimental data, in order to provide accurate driving force. Stability of the control algorithm is rigorously proved by Lyapunov methods. Experiments of full-size vehicles are conducted to validate the performance of the control approach.

This paper presents the mathematical derivation and experimental validation of a computational model, which accurately predicts static, large-strain deformations of tendon driven non-slender soft-arm manipulators subjected to gravity. The large strain behaviors are captured by employing the Green-Lagrange strain and by deriving analytical expressions for the variation of the equivalent Young modulus of the structure due to the large strains. No simplifying assumptions are made regarding the curvature of the structure, the stretching or the compression. Furthermore the paper proposes an iterative method for numerically solving the resultant non-linear system of coupled differential equations and demonstrates a number of application scenarios. The model is experimentally validated using a set-up comprising one segment of tendon driven soft-arm, which integrates stretchable and compressible hyperelastic (rubber-type) materials into its non-homogeneous back bone structure.

Understanding the behaviors of human-driven vehicles such as acceleration and braking are critical for the safety of the near-future mixed transportation systems which involve both automated and human-driven vehicles. Existing approaches in modeling human driving behaviors including driver-model-based approaches and heuristic approaches have issues in either model accuracy or scalability limitation to new situations. To address these issues, this paper proposes a new inverse model predictive control (IMPC) based approach to model longitudinal human driving behaviors. The approach incorporates the internal decision making process of humans, and achieves better predicting accuracy and improved scalability to different situations. The modeling, learning, and prediction of longitudinal human driving behaviors using the proposed IMPC approach are presented. Experimental results validate the effectiveness and advantages of the approach.

In this paper, we describe the use of calibration techniques to help improving prediction accuracy of models for soft actuators. A finite-difference model considering anisotropy of soft actuator flexible chamber is combined with experimental identification to assess actuator motion. Identifiability conditions are exploited to determine suitable experimental conditions, taking into account the manufacturing process capabilities, multimaterial additive manufacturing in this context. As a second contribution, a soft actuator with Wren mechanism for reinforcement is being developed. It is considered for rotary actuation, with the description of a proof of concept. The interest of the calibration technique and the actuator design are finally discussed.

The German Aerospace Center (DLR) has developed the Terramechanics Robotics Locomotion Lab (TROLL) to provide a feasible testing facility for developing planetary exploration rovers, as well as validating terramechanical models. A robotic manipulator is used to provide the required degrees of freedom to the mounted wheel or subsystem, making it necessary to actively control the interaction force of the wheel-soil contact. This paper is concerned with the development of a feasible force control strategy for the testbench wheel-soil contact during single-wheel experiments. For this purpose, a terramechanical model has been developed to accurately map the dynamic processes relevant for the force control design, which is later used in a testbench simulation framework to predict and evaluate the performance of control strategies. The Adaptive Admittance Control (AAC) scheme developed is adapting the gain based on the current control deviation, the rotational velocity of the wheel and an estimated soil stiffness during the experiment. The AAC is evaluated using a benchmark single-wheel experiment and shows superior performance compared to standard admittance control.

Endowing continuum robots with compliance while it interacts with the internal environment of the human body is essential to prevent damage to the robot and the surrounding tissues. Compared with passive compliance, active compliance has the advantages in terms of increasing the force transmission ability and improving safety with monitored force output. Previous studies have demonstrated that active compliance can be achieved based on a complex model of the mechanics combined with a traditional machine learning technique such as a support vector machine. This paper proposes a recurrent neural network (RNN) based approach that avoids the complexity of modeling while capturing nonlinear factors such as hysteresis, friction and delay of the electronics that are not easy to model. The approach is tested on a 3-tendon single-segment continuum robot with force sensors on each cable. Experiments are conducted to demonstrate that the continuum robot with an RNN based feed-forward controller is capable of responding to external forces quickly and entering an unknown environment compliantly.

It is becoming increasingly popular to use parallel mechanisms as modular subsystem units in the design of various robots for their superior stiffness, payload-to-weight ratio and dynamic properties. This leads to series-parallel hybrid robotic systems which pose several challenges in their modeling and control e.g. resolution of loop closure constraints, large size of their spanning tree etc. These robots are typically position-controlled and when equipped with real time dynamic control, often a simplified inverse dynamic model of these systems is utilized. However, the trade-offs of this model simplification has not been studied previously. This paper presents a representative study of the neglected dynamics by introducing some error metrics which are useful in highlighting the advantages and disadvantages of such model simplification. The study is guided with the help of a series-parallel humanoid leg which has been recently developed at DFKI-RIC.

This paper develops a new constrained model predictive control for a dynamic path tracking of an off-road mobile robot with a double steering axle. The controller is based on a dynamic model that includes wheel-ground lateral slippage and terrain geometry parameters. It is formulated as an optimization problem that computes at each time-step the optimal front and rear steering angles required to perform a desired path, with respect to multiple constraints, essentially the steering joint limits and the tire adhesion area bounds (i.e., pseudo-sliding zone limits). The capabilities of such a path tracking controller are shown and discussed through numerical simulations and experiments on a real off-road mobile robot at different speeds.

Robotic calibration allows for the fusion of data from multiple sensors such as odometers, cameras, etc., by providing appropriate relationships between the corresponding reference frames. For wheeled robots equipped with camera/lidar along with wheel encoders, calibration entails learning the motion model of the sensor or the robot in terms of the data from the encoders and generally carried out before performing tasks such as simultaneous localization and mapping (SLAM). This work puts forward a novel Gaussian Process-based non-parametric approach for calibrating wheeled robots with arbitrary or unknown drive configurations. The procedure is more general as it learns the entire sensor/robot motion model in terms of odometry measurements. Different from existing non-parametric approaches, our method relies on measurements from the onboard sensors and hence does not require the ground truth information from external motion capture systems. Alternatively, we propose a computationally efficient approach that relies on the linear approximation of the sensor motion model. Finally, we perform experiments to calibrate robots with un-modelled effects to demonstrate the accuracy, usefulness, and flexibility of the proposed approach.

This paper introduces a new mobile robot localization solution consisting of two main modules: a Particle-Filter based Localization (PFL) and a Reinforcement-Learning based map updating, integrating relative measurements and absolute indoor positioning sensor (A-IPS) data. Concerning localization using 2D-LiDARs, featureless areas are known to be problematic. To solve this problem a classic PFL approach was modified to incorporate A-IPS position measurements in the prediction and update stages. The localization approach has the particularity of including the possibility of updating the map whenever major modifications are detected in the environment in relation to the current localization map. Due to the random sampling-based nature of the PFL, an associated map update solution is not trivial since small inconsistencies in the estimated pose can lead to erroneous map associations. The proposed method learns to decide by assigning higher rewards the greater is the overlap between the map and the 2DLIDAR scans, via RL, and then a proper update of the map is achieved. Validation of the proposed pipeline was carried out in a differential drive platform with algorithms developed in ROS. Tests were performed in two scenarios in order to assess the performance of both the localization module and the map update stage. The results show that the proposed localization method offers improvements in relation to known approaches, and consequently suggest promising perspectives for the proposed map update decision framework.

Learning by imitation is a powerful way that can reduce the complexly in searching space. It could help the mobile robot to acquire new skills from interaction with a human-being in natural way. In this paper, the dynamic movement primitives (DMPs) is utilized to imitate the trajectory from human walking. DMPs is a modified formulation of virtual spring-dampers (VSD) system that enjoys better fitting performance in learning. Further, while dealing with the trajectory tracking problem of mobile robots, a novel nonlinear model predictive control (MPC) approach is proposed for motion control. The nonlinear MPC scheme applies a new neural network named Varying-parameter Lagrangian Neural Network (VP-LNN) to solve a Quadratic Programming (QP) problem by iterating over a finite receding horizon. The new network of VP-LNN can converge to the global optimal solution. Thus, a new human-robot interaction (HRI) scheme for mobile robot is proposed, which can reduce the complexity in motion planning in various applications.