ICRA 2017 Papers

In this paper, a sliding-window two-camera vision-aided inertial navigation system (VINS) is presented in the square-root inverse domain. The performance of the system is assessed for the cases where feature matches across the two-camera images are processed with or without any stereo constraints (i.e., stereo vs. binocular). To support the comparison results, a theoretical analysis on the information gain when transitioning from binocular to stereo is also presented. Additionally, the advantage of using a two-camera (both stereo and binocular) system over a monocular VINS is assessed. Furthermore, the impact on the achieved accuracy of different image-processing frontends and estimator design choices is quantified. Finally, a thorough evaluation of the algorithm's processing requirements, which runs in real-time on a mobile processor, as well as its achieved accuracy as compared to alternative approaches is provided, for various scenes and motion profiles.

This paper proposes a driver assistance system with a collaborative control framework between the human driver and a Collision Avoidance System (CAS) for vehicles on the road while considering the driver drowsiness status. Driver drowsiness detection is performed with two inputs: driver's facial data from a camera mounted in front of the driver and steering wheel data from the car controller system. We use the driver's drowsiness state as an input to the collaborative control framework in which the CAS algorithm runs in parallel with the human control and only intervenes under certain situations to assist the human driver. Experiments were performed on a simulated vehicle driving testbed to evaluate our drowsiness detection system and demonstrate the effectiveness of the proposed collaborative control framework.

Robotic agents that do everyday manipulation tasks can hugely benefit from being able to predict consequences of their actions just before the execution. However, such a simulation technique is usually computationally-expensive and may not be achieved with agents' self computing power. For this problem, cloud robotics may offer a solution. Cloud robotics is an emerging field in the intersection of robotics and cloud computing which enables robots to access a greater amount of processing power and storage capacity than it can employ within itself. In this work, we introduce a mental simulation service to, one of the cloud engines, openEASE [1]. Using this service, researchers and robots can describe the world model, the state of the agent and the problem that is being dealt with. In return, it simulates the world and runs a learning algorithm and suggests a solution how the robotic agent can handle the problem. This service does not only offer a free remote access to simulation which is computationally expensive but also thanks to OPEnEASE's rich reasoning techniques these simulated experiments can be reasoned later on using prolog queries.

In support of Cloud Robotics, Robotics and Automation as a Service (RAaaS) frameworks have the potential to reduce the complexity of software development, simplify software installation and maintenance, and facilitate data sharing for machine learning. In this proof-of-concept paper, we describe Berkeley Robotics and Automation as a Service (Brass), a RAaaS prototype that allows robots to access a remote server that hosts a robust grasp-planning system (Dex-Net 1.0) that maintains data on hundreds of candidate grasps on thousands of 3D object meshes and uses perturbation sampling to estimate and update a stochastic robustness metric for each grasp. Results suggest that such a system can increase grasp reliability over naive locally-computed grasping strategies with network latencies of 30 and 200 msec for servers 500 and 6000 miles away, respectively. We also study how the system can use execution reports from robots in the field to update grasp recommendations over time.

There is currently a scarcity of wearable robotic devices that can practically provide physical assistance in a range of real world activities. Soft wearable exoskeletons, or exosuits, have the potential to be more portable and less restrictive than their rigid counterparts. In this paper, we present the design of an actuation system that has been optimized for use in a soft exosuit for the human arm. The selected design comprises a DC motor and gearbox, a flexible cable conduit transmission, and a custom series elastic force sensor. Placed in series with the transmission conduit, the custom compliant force sensor consists of a translational steel compression spring with a pair of Hall effect sensors for measuring deflection. The custom sensor is validated as an accurate means of measuring cable tension, and it is shown that it can be used in feedback to control the cable tension with high bandwidth. The dynamic effect of the cable-conduit transmission on the force felt at the user interface is characterized by backdriving the system as it renders a range of virtual impedances to the user. We conclude with recommendations for the integration of such an actuation system into a full wearable exosuit.

We address the problem of camera-to-laserscanner calibration using a checkerboard and multiple imagelaser scan pairs. Distinguishing which laser points measure the checkerboard and which lie on the background is essential to any such system. We formulate the checkerboard extraction as a combinatorial optimization problem with a clear cut objective function. We propose a branch-and-bound technique that deterministically and globally optimizes the objective. Unlike what is available in the literature, the proposed method is not heuristic and does not require assumptions such as constraints on the background or relying on discontinuity of the range measurements to partition the data into line segments. The proposed approach is generic and can be applied to both 3D or 2D laser scanners as well as the cases where multiple checkerboards are present. We demonstrate the effectiveness of the proposed approach by providing numerical simulations as well as experimental results.

Confocal endomicroscopy can be used for identification of early mucosal dysplasia in various gastrointestinal conditions, and has particular potential in the monitoring of Barrett's oesophagus and the early stages of oesophageal cancer. However, it can be difficult to systematically scan a significant area of the oesophagus because of the small field-of-view and limited flexibility of the probe. Tissue deformation and inconsistent probe-tissue contact also make it difficult to form large mosaics. A mechanical scanning device is therefore desirable for controlled, large area surface scanning and mosaicing of the oesophagus. This paper proposes a robotic catheter encapsulated in an inflatable balloon, providing stable scanning over the oesophageal surface. It has an outer diameter of 3 mm, making it suitable for deployment through an endoscope working channel, and uses a custom endomicroscopy probe based on a leached flexible fibre bundle and an external confocal laser scanning system. Detailed mechanical performance and image quality evaluations were performed to assess the clinical potential of the device. In ex vivo studies using swine oesophagus, long helical scans were obtained, demonstrating that the device is able to scan the lumen stably and maintain good probe-tissue contact. The experimental results demonstrate the potential of the robotic catheter for systematic high-resolution imaging of the oesophageal mucosa, potentially reducing or even eliminating the need for physical biopsy.

We present the Model-Predictive Motion Planner (MPMP) of the Intelligent Autonomous Robotic Automobile (IARA). IARA is a fully autonomous car that uses a path planner to compute a path from its current position to the desired destination. Using this path, the current position, a goal in the path and a map, IARA's MPMP is able to compute smooth trajectories from its current position to the goal in less than 50 ms. MPMP computes the poses of these trajectories so that they follow the path closely and, at the same time, are at a safe distance of occasional obstacles. Our experiments have shown that MPMP is able to compute trajectories that precisely follow a path produced by a Human driver (distance of 0.15 m in average) while smoothly driving IARA at speeds of up to 32.4 km/h (9 m/s).

A 3 DoF parallel cable driven body weight support (BWS) system has been developed for the University of Utah's Treadport Locomotion Interface, for purposes of rehabilitation, simulation of steep slopes, and display of reduced gravity environments. The Treadport's large belt (6 by 10 feet) requires a multi-cable support system to ensure that the unloading forces are close to vertical. This paper presents the design and experimental validation, including the system model and force control.

This paper presents the design, control, and experimental validation of a novel fully-actuated aerial robot for physically interactive tasks, named Tilt-Hex. We show how the Tilt-Hex, a tilted-propeller hexarotor is able to control the full pose (position and orientation independently) using a geometric control, and to exert a full-wrench (force and torque independently) with a rigidly attached end-effector using an admittance control paradigm. An outer loop control governs the desired admittance behavior and an inner loop based on geometric control ensures pose tracking. The interaction forces are estimated by a momentum based observer. Control and observation are made possible by a precise control and measurement of the speed of each propeller. An extensive experimental campaign shows that the Tilt-Hex is able to outperform the classical underactuated multi-rotors in terms of stability, accuracy and dexterity and represent one of the best choice at date for tasks requiring aerial physical interaction.

This paper presents a novel approach to estimating the continuous six degree of freedom (6-DoF) pose (3D translation and rotation) of an object from a single RGB image. The approach combines semantic keypoints predicted by a convolutional network (convnet) with a deformable shape model. Unlike prior work, we are agnostic to whether the object is textured or textureless, as the convnet learns the optimal representation from the available training image data. Furthermore, the approach can be applied to instance- and class-based pose recovery. Empirically, we show that the proposed approach can accurately recover the 6-DoF object pose for both instance- and class-based scenarios with a cluttered background. For class-based object pose estimation, state-of-the-art accuracy is shown on the large-scale PASCAL3D+ dataset.

Autonomous crop monitoring at high spatial and temporal resolution is a critical problem in precision agriculture. While Structure from Motion and Multi-View Stereo algorithms can finely reconstruct the 3D structure of a field with low-cost image sensors, these algorithms fail to capture the dynamic nature of continuously growing crops. In this paper we propose a 4D reconstruction approach to crop monitoring, which employs a spatio-temporal model of dynamic scenes that is useful for precision agriculture applications. Additionally, we provide a robust data association algorithm to address the problem of large appearance changes due to scenes being viewed from different angles at different points in time, which is critical to achieving 4D reconstruction. Finally, we collected a high-quality dataset with ground-truth statistics to evaluate the performance of our method. We demonstrate that our 4D reconstruction approach provides models that are qualitatively correct with respect to visual appearance and quantitatively accurate when measured against the ground truth geometric properties of the monitored crops.

This paper describes an intuitive method for the design and fabrication of small-scale robots with multi-material compliant mechanisms in 3D. The rigid components are 3D printed, and flexures are inserted into the rigid components, creating the final mechanism. The assembled mechanisms are robust, requiring over 1 N of force to delaminate, and surviving 150,000 cycles of bending without failure. A 6 g walking quadrupedal millirobot is presented as a case study for the design and manufacturing methodology. The quadrupedal millirobot has been demonstrated moving at top speeds of 9 mm/s (0.3 body lengths/s).

This paper presents the design, fabrication, modeling, and performance characterization of a new 3D-printed ionic polymer-metal composite (IPMC) soft crawling robot. First, a precursor to an ionomeric polymer material (Nafion) is used to 3D print modular leg and body sections to create a caterpillar-like robot. Then, the printed components are activated, plated with electrodes, assembled together to create the robot from smart electroactive polymer material. The 3D-printed robot exploits the unique capabilities of IPMC materials; specifically, the actuation of the hydrated 3D-printed leg and body sections can be controlled by applying a voltage signal. In particular, the IPMC legs can grip and the body sections can expand and contract, allowing the robot to propel itself forward and backward. The newly-developed 3D-printing process and the robot design and modeling process are described. Experimental results are presented that show the prototype robot moving along a tube like a caterpillar or inchworm.

Current self-driving car systems operate well in sunny weather but struggle in adverse conditions. One of the most commonly encountered adverse conditions involves water on the road caused by rain, sleet, melting snow or flooding. While some advances have been made in using conventional RGB camera and LIDAR technology for detecting water hazards, other sources of information such as polarization offer a promising and potentially superior approach to this problem in terms of performance and cost. In this paper, we present a novel stereo-polarization system for detecting and tracking water hazards based on polarization and color variation of reflected light, with consideration of the effect of polarized light from sky as function of reflection and azimuth angles. To evaluate this system, we present a new large `water on road' datasets spanning approximately 2 km of driving in various on-road and off-road conditions and demonstrate for the first time reliable water detection and tracking over a wide range of realistic car driving water conditions using polarized vision as the primary sensing modality. Our system successfully detects water hazards up to more than 100m. Finally, we discuss several interesting challenges and propose future research directions for further improving robust autonomous car perception in hazardous wet conditions using polarization sensors.

Soft robots have recently demonstrated impressive abilities to adapt to objects and their environment with limited sensing and actuation. However, mobile soft robots are typically fabricated using laborious molding processes that result in limited actuated degrees of freedom and hence limited locomotion capabilities. In this paper, we present a 3D printed robot with bellowed soft legs capable of rotation about two axes. This allows our robot to navigate rough terrain that previously posed a significant challenge to soft robots. We present models and FEM simulations for the soft leg modules and predict the robot locomotion capabilities. We use finite element analysis to simulate the actuation characteristics of these modules. We then compared the analytical and computational results to experimental results with a tethered prototype. The experimental soft robot is capable of lifting its legs 5.3 cm off the ground and is able to walk at speeds up to 20 mm/s (0.13 bl/s). This work represents a practical approach to the design and fabrication of functional mobile soft robots.

This paper proposes a nonprehensile manipulation scheme using the vibration of a plate, where a 3-DoF (degree of freedom) motion of a part is controlled using a single actuator. A manipulator is introduced, the end effector of which is a flat plate. The manipulator employs a hybrid joint mechanism with a twisted axis layout. The characteristic of this mechanism is that the shape of the trajectory of the plate can be varied three dimensionally based on the sinusoidal displacement input to the actuator. The behaviors of multiple point masses on the plate are then analyzed to understand the approximated 3-DoF motion of a part. The results of the simulation reveal that the whirlpool-like characteristics of the trajectory map, which aids in the rotational and translational motions of the part, can be governed. Based on the characteristics, five primitives for manipulating the part are defined and their arrangements for a 3-DoF parts feeding task are shown. Finally, the proposed approach is demonstrated via experiments.