ICRA 2015 论文列表

Terrain compliance is a critical parameter for the performance of legged locomotion. In this work, a single actuator monopod robot hopping on rough compliant terrain is considered. Based on our controller for flat compliant terrains, this paper introduces the necessary modifications, which allow the robot to tackle the disturbance of small inclinations. Using the developed method, the robot is examined on its performance to traverse rough terrains, while maintaining the goals of reaching a desired height and forward velocity. As the increased compliance and inclination alter the energy requirements from the controller actuator, the Cost of Transport (CoT) index for a number of scenarios is studied. The correlation between terrain parameters and the CoT is presented, and useful conclusions, which can aid the understanding of the behavior of legged robots in realistic terrains are extracted.

In this paper, we propose a resource-efficient system for real-time 3D terrain reconstruction and landing-spot detection for micro aerial vehicles. The system runs on an on-board smartphone processor and requires only the input of a single downlooking camera and an inertial measurement unit. We generate a two-dimensional elevation map that is probabilistic, of fixed size, and robot-centric, thus, always covering the area immediately underneath the robot. The elevation map is continuously updated at a rate of 1 Hz with depth maps that are triangulated from multiple views using recursive Bayesian estimation. To highlight the usefulness of the proposed mapping framework for autonomous navigation of micro aerial vehicles, we successfully demonstrate fully autonomous landing including landing-spot detection in real-world experiments.

To provide the elderly with appropriate and timely caregiving in activities of daily life (ADL), it is desired for robots to have the capability of intention awareness (IA). Different from existing context-specific intention awareness (CSIA) approaches which are based on a limited and passive knowledge database, our approach, named `WebIA', adopts a novel web query approach to establish a vast and active knowledge database for robots to perform CSIA. In our method, robots are endowed with comprehensive commonsense knowledge towards the correlations of the surrounding environment and human intentions. WebIA enables robots to effectively infer intentions in both trained familiar and untrained new situations. By performing several experiments, we evaluated two aspects of the WebIA approach: the effectiveness of IA in familiar situations and the self-learning ability in new situations.

This paper presents the contact-consistent elastic strips (CES) framework, a motion planning approach capable of producing complex multi-contact whole-body humanoid behaviors in dynamic environments. Planning multi-contact motions for humanoid robots is known to be non-trivial since it involves aspects of autonomous balancing, obstacle avoidance, ensuring global connectivity of the workspace and concurrent consideration of the kinematic and dynamic constraints. Previous works at motion planning for humanoid systems tend to focus on joint space planning and deal with obstacle avoidance and contact-point searching as the separated problems. CES framework, however, simultaneously considers all these requirements and constraints in task space while planning a valid sequence of contact-points and corresponding motions. This resulted in considerable improvements to efficiency and significantly reduced planning time. With the use of CES framework, complex multi-contact locomotion behaviors and real-time adjustments to the robot motions in 3D unstructured environments is possible. Several simulations are demonstrated to evaluate and verify the performance of CES framework.

We study the problem of building a sensor model for the purpose of simulation. Our work is motivated by the potential impact of realistic simulators on the development cycle of software for real robots. The case is made for building models from approximate state information, relieving the burden of ground truth. Unlike calibration, where the goal is to identify and remove error from a signal, our aim to reproduce the signal in its entirety, including its error properties. Instead of physically modeling sensor behavior, a data-driven approach is taken. The implementation of our approach to simulate a simple but noisy laser rangefinder is described. Finally, approaches to validate the simulator are discussed. We compare not only raw sensor predictions, but also overall performance of algorithms on simulated versus real data.

The cooperative path planning problem seeks to determine a path for a group of robots which form temporary teams to perform tasks that require multiple robots. The multi-scale effects of simultaneously coordinating many robots distributed across the workspace while also tightly coordinating robots in cooperative teams increases the difficulty of planning. This paper describes a new approach to cooperative path planning called Constraint Manifold Subsearch (CMS). CMS builds upon M*, a high performance multirobot path planning algorithm, by modifying the search space to restrict teams of robots performing a task to the constraint manifold of the task. CMS can find optimal solutions to the cooperative path planning problem, or near optimal solutions to problems involving large numbers of robots.

Vision-aided localization systems are often utilized in urban settings to take advantage of structured environment, high availability of unique visual features, as well as complimenting aiding measurements from Global Navigation Satellite System (GNSS). In this paper, we present a case study for roadway texture mapping that combines low-cost sensor measurements that are already available on many production vehicles (e.g. single frequency GPS, wheel odometry, and a forward looking camera) together with a satellite image. The aim of the method presented here is to obtain high resolution texture of the ground plane that is consistent with the low-resolution satellite image through an optimization process that estimates the smooth vehicle trajectory using Maximum-a-Posteriori (MAP). The main benefit of this system comes from the facts that: (1) it utilizes only low-cost sensors and information that are readily available, (2) it can be easily embedded into existing maps. Data and analysis of a drive captured around a block is used in this study.

Robust control maintains stability and performance for a fixed amount of model uncertainty but can be conservative since the model is not updated online. Learning-based control, on the other hand, uses data to improve the model over time but is not typically guaranteed to be robust throughout the process. This paper proposes a novel combination of both ideas: a robust Min-Max Learning-Based Nonlinear Model Predictive Control (MM-LB-NMPC) algorithm. Based on an existing LB-NMPC algorithm, we present an efficient and robust extension, altering the NMPC performance objective to optimize for the worst-case scenario. The algorithm uses a simple a priori vehicle model and a learned disturbance model. Disturbances are modelled as a Gaussian Process (GP) based on experience collected during previous trials as a function of system state, input, and other relevant variables. Nominal state sequences are predicted using an Unscented Transform and worst-case scenarios are defined as sequences bounding the 3σ confidence region. Localization for the controller is provided by an on-board, vision-based mapping and navigation system enabling operation in large-scale, GPS-denied environments. The paper presents experimental results from testing on a 50 kg skid-steered robot executing a path-tracking task. The results show reductions in maximum lateral and heading path-tracking errors by up to 30% and a clear transition from robust control when the model uncertainty is high to optimal control when model uncertainty is reduced.

Ensuring safety in partially-known environments is a critical problem in robotics since the environment is perceived through sensors and the environment cannot be completely known ahead of time. Prior work has considered the problem of finding positive control invariant sets (PCIS). However, this approach limits the planning horizon of the motion planner since the PCIS must lie completely in the limited known part of the environment. Here we consider the problem of guaranteeing safety by ensuring the existence of at least one PCIS in partially-known environments leading to an extension of the PCIS concept. It is shown, that this novel method is less conservative than the common PCIS approach and robust to unknown small obstacles which might appear in the close vicinity of the robot. An example implementation for loiter circles and power line obstacles is presented. Simulation scenarios are used for validating the proposed concept.

This paper presents a framework to compute stable contact interfaces for automatically designing customized jigs used in bone surgeries. Given the surface model of a bone represented by polygonal meshes, we find out a small region on the surface to be used as the interface of a customized jig so that it can be stably fixed on the bone under a directional clamping force. The variation of directions on the clamping force is allowed in our formulation. Moreover, the surface region serves as the interface of stable contact must also be disassemblable so that the jig and the bone can be separated after removing the clamping force. The analysis of stable contact is formulated on a Gaussian map by the common regions of half-spaces according to the motion restrictions. A flooding algorithm is proposed to determine those disassemblable and stable contact interfaces on the surface of a bone, where the contact surfaces are later converted into the solid model of a jig to be fabricated by additive manufacturing. Experimental tests are taken to verify the stable contact between a bone and the jig generated by our approach.

Weakly electric fish use a navigational technique called electrolocation to investigate their surroundings for predator, prey and obstacles. Obstacles and other global stimuli are perceived as perturbations to the fish's self-generated electric field, which provide relevant navigational cues to the fish. In this work a control strategy based on electrolocation for performing obstacle avoidance in electrically heterogeneous environments is presented and validated. The control strategy developed for a straight tunnel is shown to be robust to small variations in the tunnel width. This justifies its application to other heterogeneous corridor-like environments. Static output feedback control is shown to achieve the desired goal of reflexive obstacle avoidance in such environments in simulation and experimentation. The proposed approach is computationally inexpensive and readily implementable on a small scale underwater vehicle, making underwater autonomous navigation feasible in real-time.

This paper presents a compliant locomotion framework for torque-controlled humanoids using model-based whole-body control. In order to stabilize the centroidal dynamics during locomotion, we compute linear momentum rate of change objectives using a novel time-varying controller for the Divergent Component of Motion (DCM). Task-space objectives, including the desired momentum rate of change, are tracked using an efficient quadratic program formulation that computes optimal joint torque setpoints given frictional contact constraints and joint position / torque limits. In order to validate the effectiveness of the proposed approach, we demonstrate push recovery and compliant walking using THOR, a 34 DOF humanoid with series elastic actuation. We discuss details leading to the successful implementation of optimization-based whole-body control on our hardware platform, including the design of a “simple” joint impedance controller that introduces inner-loop velocity feedback into the actuator force controller.

Nowadays sampling-based motion planners use the power of randomization to compute multidimensional motions at high performance. Nevertheless the performance is based on problem-dependent parameters like the weighting of translation versus rotation and the planning range of the algorithm. Former work uses constant user-adjusted values for these parameters which are defined a priori. Our new approach extends the power of randomization by varying the parameters randomly during runtime. This avoids a preprocessing step to adjust parameters and moreover improves the performance in comparison to existing methods in the majority of the benchmarks. Our method is simple to understand and implement. In order to compare our approach we present a comprehensive experimental analysis about the parameters and the resulting performance. The algorithms and data structures were implemented in our own library RASAND, but we also compare the results of our work with OMPL [12] and the commercial software Kineo™ Kite Lab [15].

Atomic force microscopy (AFM) is a powerful measurement instrument, which has been widely implemented in various fields. To enhance the maneuverability, an AFM can be modified and its cantilever can be controlled as a robotic end-effector. To precisely control the nanorobots, their scanners inherent hysteresis, which is the main disadvantage, should be compensated sufficiently. Mostly, hysteresis compensators used in AFM control system only employ the symmetric models, which cannot well represent the asymmetric hysteresis phenomena of piezo-scanners. However, under many cases, the smart material based scanners possess asymmetric hysteresis slightly or seriously, which is far more complex than the regular symmetric case; in addition, for commercial or customized AFMs, the scanners are usually controlled without feedback. These drawbacks tackle further improvement of positioning accuracy of AFM systems. To effectively describe and further reduce the hysteresis of general cases, we propose a new type of generalized Prandtl-Ishlinskii (PI) operator based superposition model, named unparallel PI (UPI) model. The flexible edge of the UPI operator can be tilted freely within a defined range, which enables it to capture asymmetric hysteresis efficiently. To cancel the hysteresis effect in an open-loop system, three inverse compensation approaches are proposed, compared, and the corresponding stability is analyzed. Experiments with AFM based nanorobot verified the proposed approaches with convincing performance.

This paper presents an overview and comparison of minimal and complete rigid body motion trajectory descriptors, usable in applications like motion recognition and programming by demonstration. Motion trajectory descriptors are able to deal with potentially unwanted variations acting on the motion trajectory such as changes in the execution time, the motion's starting position, or the viewpoint from which the motion is observed. A suitable rigid body motion trajectory descriptor retains only the trajectory information relevant to the application. This paper compares different trajectory descriptors for rigid body motion and validates their usefulness for dealing with motion variation in a motion recognition experiment. Furthermore, a new type of invariant trajectory descriptor is introduced based on the Frenet-Serret formulas.

This work provides a case study of both fixed-wing and rotorcraft small unmanned aerial systems (SUAS) used in a deployment at the SR530 mudslides in Washington state and compares the types of flight paths used by each vehicle type. Previously aerial imagery from SUAS have produced 2D and 3D reconstructions of simple terrain, but have not been used in complex terrain which encompasses both flat areas and drastic changes in the height of ground level, such as a mudslide. In this deployment, both types of SUAS platforms were used to collect imagery over terrain varied nearly 200m in elevation but different paths were used due to the complexity of the terrain, safety, privacy, and platform-specific limitations. The deployment found that paths with fixed-wing platforms can be thought of as stacked horizontal planes while rotorcraft can cover complex terrain with a set of vertical planes. The different paths contribute to autonomous path planning, particularly to accommodate vertical planes, and to general understanding of how different SUAS can be applied to challenging terrains. Future work in path planning should incorporate Geographic Information Systems (GIS) information to facilitate flight paths in vertical planes and to maintain altitude restrictions relative to radically changing elevations of a landscape.

Discrete-time estimation of rigid body attitude and angular velocity without any knowledge of the attitude dynamics model, is treated using the discrete Lagrange-d'Alembert principle. Using body-fixed sensor measurements of direction vectors and angular velocity, a Lagrangian is obtained as the difference between a kinetic energy-like term that is quadratic in the angular velocity estimation error, and an artificial potential obtained from Wahba's function. An additional dissipation term that depends linearly on the angular velocity estimation error is introduced, and the discrete Lagrange-d'Alembert principle is applied to the Lagrangian with this dissipation. An implicit and an explicit first-order version of this discrete-time estimation scheme is presented. A comparison of this estimator is made with certain state-of-the-art attitude estimators in the absence of bias in sensor readings. Numerical simulations show that this estimator is robust and unlike extended Kalman filter-based schemes, its convergence does not depend on the gain values. In addition, the variational estimator is found to be more computationally efficient than these other estimators.

We present Lynx-robot, a quadruped, modular, compliant machine. It alternately features a directly actuated, single-joint spine design, or an actively supported, passive compliant, multi-joint spine configuration. Both spine configurations bend in the sagittal plane. This study aims at characterizing these two, largely different spine concepts, for a bounding gait of a robot with a three segmented, pantograph leg design. An earlier, similar-sized, bounding, quadruped robot named Bobcat with a two-segment leg design and a directly actuated, single-joint spine design serves as a comparison robot, to study and compare the effect of the leg design on speed, while keeping the spine design fixed. Both proposed spine designs (single rotatory and active and multi-joint compliant) reach moderate, self-stable speeds.

Multi-robot deployments have the potential for completing tasks more efficiently. For example, in simultaneous localization and mapping (SLAM), robots can better localize themselves and the map if they can share measurements of each other (direct encounters) and of commonly observed parts of the map (indirect encounters). However, performance is contingent on the quality of the communications channel. In the underwater scenario, communicating over any appreciable distance is achieved using acoustics which is low-bandwidth, slow, and unreliable, making cooperative operations very challenging. In this paper, we present a framework for cooperative SLAM (C-SLAM) for multiple autonomous underwater vehicles (AUVs) communicating only through acoustics. We develop a novel graph-based C-SLAM algorithm that is able to (optimally) generate communication packets whose size scales linearly with the number of observed features since the last successful transmission, constantly with the number of vehicles in the collective, and does not grow with time even the case of dropped packets, which are common. As a result, AUVs can bound their localization error without the need for pre-installed beacons or surfacing for GPS fixes during navigation, leading to significant reduction in time required to complete missions. The proposed algorithm is validated through realistic marine vehicle and acoustic communication simulations.