IROS 2017 Papers

In running, legged animals take advantages of their flexible bodies to increase their locomotion performance, resulting in superior mobility and fast motions. Previous works on quadruped robots mostly used the simplified model with massless legs. However, in reality, a leg consists of two or three segments with mass and inertia. We study herein the bounding of quadruped robots with a model closer to a real animal, i.e., a model having articulated legs with joints with stiffness and damping properties, and a passive spinal joint, to obtain an insight into the robot's dynamic behaviors. The effects of spine flexibility were investigated on quadrupedal bounding gait by changing dynamic properties and hardware parameters and then comparing the performance of models with and without a spinal joint. The model comparisons revealed that body flexibility significantly increased stride length and strongly affected the dynamic of the body and the energy efficiency.

Manta ray's pectoral fins have been a great source of inspiration for propulsive mechanism of autonomous underwater vehicles, due to their propulsive capability. The geometry (shape) and flexibility factors of these fins have been hypothesized to be determinants of the propulsive capability of the fins in terms of thrust generation. In particular, the sweep angle factor has been omitted from previous studies, where it has been commonly set to about 30 degrees. This paper investigates the effects of sweep angle on thrust generation of oscillatory pectoral fins. Forty different fins were designed and fabricated to be experimented in a water channel, which involved measurement of thrust generated by the fins. The experiment was conducted under free stream (0.5 m/s) and still water conditions. Five different sweep angles (0, 10, 20, 30, 40 degrees) were incorporated into eight base designs of different flexibility characteristics to make up the 40 fins. To consider only sweep angle, other geometrical factors were not varied. Within the range of the sweep angle considered, the experimental results showed that sweep angle has no significant influence on the fins' thrust generation, under both free stream and still water conditions. Overall, it can be concluded that sweep angle may not be a determinant of oscillatory pectoral fins' thrust generation.

One of the main abilities that the robots need to maintain is to efficiently communicate with people in a humanly manner. Thus, human activity recognition (HAR) would be an integral part of such a human-robot interaction system. One of the major challenges in HAR is that the individuals perform their activities in different manners. Furthermore, there is a very wide range of different types of activities that the robots would require to understand. Some activities are simple, quick and short (e.g., sit down), while many others are complex, have many details and span through a long range of time (e.g., wearing contact lens). In this paper, we model the recognition of activities into a sequence-labeling problem and propose a new probabilistic graphical model (PGM) that can recognize both short/long-range activities, by introducing a hierarchical classification model and including extra links and loopy conditions in our PGM. To optimize the PGM and obtain its parameters during training, we use a structured prediction technique, a general framework that involves latent structured support vector machines (LSSVM) and hidden-state conditional random fields (HCRF). We evaluate our method on two widely used datasets (CAD-60 & UT-Kinect) that contain both activity types. Our obtained results are promising and show that our method can recognize both types of activities effectively, while most of the previous works only focused on one of these two major types. We further explore distributed processing techniques, since our method can easily be distributed over processing nodes. We also propose an efficient divide-and-merge technique to further speedup the training step.

This paper describes a structured light-based sensor for hazard avoidance in planetary environments. The system presented here can also be used in terrestrial applications constrained by reduced onboard power and computational complexity and low illumination conditions. The sensor consists on a calibrated camera and laser dot projector system. The onboard hazard avoidance system determines the position of the projected dots in the image and through a triangulation process detects potential hazards. The paper presents the design parameters for this sensor and describes the image based solution for hazard avoidance. The system presented here was tested extensively in day and night conditions in Lunar analogue environments. The current system achieves over 97% detection rate with 1.7% false alarms over 2000 images.

Autonomous robotic assembly by mobile field robots has seen significant advances in recent decades, yet practicality remains elusive. Identified challenges include better use of state estimation to and reasoning with uncertainty, spreading out tasks to specialized robots, and implementing representative joining methods. This paper proposes replacing 1) self-correcting mechanical linkages with generalized joints for improved applicability, 2) assembly serial manipulators with parallel manipulators for higher precision and stability, and 3) all-in-one robots with a heterogeneous team of specialized robots for agent simplicity. This paper then describes a general assembly algorithm utilizing state estimation. Finally, these concepts are tested in the context of solar array assembly, requiring a team of robots to assemble, bond, and deploy a set of solar panel mockups to a backbone truss to an accuracy not built into the parts. This paper presents the results of these tests.

This paper presents a robot-assisted system to obtain elastic information of a moving tissue using a 2-D ultrasound probe actuated by a 6 degrees of freedom robotic arm. The proposed method combines ultrasound image-based visual servoing, force control and non-rigid motion estimation. We present how the motion estimation, while the force control and visual sevoing are enabled, is useful to compute the strain map of the tissue. Ex-vivo experiments performed on a moving abdominal phantom demonstrate the efficiency and robustness of this methodology to compute the strain information of a tissue in motion.

This paper presents a control method for steering three dimensional (3D) dynamically walking bipeds that are engaged in cooperative tasks such as object transportation. Towards achieving safe interaction with a leading human (or robot) collaborator, the walking biped is required to exhibit compliance at the port of interaction, while simultaneously adapting its walking pattern in response to the perceived interaction forces. To address these issues, we propose a method that fuses impedance control of the biped's arm with position control of its legs in a way that the biped adaptively modifies its stepping pattern according to the collaborator's intentions. The method is applied on a 3D bipedal robot that is driven in the workspace by a collaborator, whose intention is communicated to the biped through the interaction force.

This paper discusses a method for generating an underactuated bipedal gait completing in one step in the presence of an upper body that behaves as zero dynamics. First, we introduce a model of a 3-link planar biped robot with an upper body, and develop the equation of motion and holonomic constraint conditions. We also describe a straightforward derivation of the relationship between the angular momentum and whole center of mass. Second, we design two control laws for achieving stable stealth walking. During the single-limb support phase, the stance- and swing-leg angles are strictly controlled to follow the desired trajectories so that the swing foot lands on the ground stealthily at the end of this phase. During the double-limb support phase, the angular position and velocity of the upper body are controlled to follow the desired trajectory for returning them to the initial state. Through numerical simulations, the fundamental gait properties and change tendencies are investigated.

Human foot injury due to being run over by a robot is a hazardous event caused by robot motion. This study aims to provide information for the safety design of robots in terms of the risk of foot fracture. A model for scaling the fracture loads of alternate specimens (bear metatarsals) to those of human metatarsals considering the effects of soft tissue is proposed and validated. Using the model, fracture tolerance of metatarsals of a human female foot of statistically typical size is predicted to be 1.24 kN at 10% fracture probability, and 0.53 kN at 1% fracture probability with 95% confidence.

This paper presents the control strategies used to adapt the actively articulated suspension system of the rover SherpaTT to irregular terrain. Experimental validation of the approach with the physical system is conducted and presented. The coordinated control of the legs constituting the suspension system is encapsulated in a Ground Adaption Process (GAP) that operates independently from high level motion commands. The GAP makes use of force and orientation measurements to control the suspension system with 20 active degrees of freedom. The active suspension is used to achieve multi-objective terrain adaption encompassing (i) active force distribution at the wheel-ground contact points, (ii) keeping all wheels in permanent ground contact, and (iii) body orientation w.r.t. gravity.

This paper presents the analytic solution of a fundamental open problem in the framework of state estimation/nonlinear observability, which is the Unknown Input Observability problem (UIO problem). The problem consists in deriving the analytic criterion that allows us to automatically obtain the state observability in presence of disturbances (or unknown inputs). In other words, the problem is to extend the well known observability rank condition to the case when the dynamics are also driven by unknown inputs. Enunciated in the seventies by the control theory community, this problem was only solved in the linear case. The solution here provided holds for nonlinear systems in presence of a single unknown input. The first part of the paper presents this analytic solution. Very surprisingly, the complexity of the overall analytic criterion is comparable to the complexity of the observability rank condition. The second part of the paper applies this analytic criterion to a robotics system when its dynamics are affected by an external disturbance (e.g., due to the presence of wind). To corroborate the results of our observability analysis we perform extensive simulations and we show that, a simple estimator based on an Extended Kalman Filter, provides results that agree with what we could expect from the observability analysis.

To enhance the robotic manipulation of deformable objects, a robust state estimator is proposed to track the object configuration in real time. A Gaussian mixture model (GMM) is constructed to register the object nodes towards the noisy point cloud. To deal with occlusion, the coherent point drift (CPD) regularization is applied on the mixture model, so as to maintain the topological structure from previous sequences of data and to infer the object states in occluded area. The state estimation is further refined by running a dynamic simulation in parallel, which guarantees the estimates to satisfy the object's physical constraints. A series of rope tracking experiments are performed to evaluate the proposed state estimator. It is shown that the object can be tracked robustly with sensor noise, outliers and massive occlusion.

Posture control is one of the fundamental requirements for legged robots to achieve mobility on uneven terrain. Most of the research on posture control is conducted on custom designed robot platforms, which are much more expensive than their commercially available counterparts. This paper presents the implementation of a standing posture control system on a low-cost, commercially available hexapod robot platform. The Virtual Model Control (VMC) strategy is implemented, along with a simple method for foot force distribution and a direct force control method that can be used with position control of low-cost joint actuators. The controller was experimentally tested on level and uneven terrain, as well as on a dynamic balance board. Ground truth measurements of the tests revealed that the results compared favorably to those obtained on more expensive, custom designed platforms.

Place recognition is an essential part of robot localization and mapping problems. Using lower data-rate sensors like 2D scanning laser rangefinders enables the robots to use less memory and computation in building maps. However, place recognition by a vehicle with 6-DOF dynamics like a quadrotor in unstructured, 3D environments like forests is challenging, especially with a sensor that only measures a planar slice of the environment. This paper extends the 2D geometry-based place recognition system of [1] to a challenging forest envirnoment with a novel procedure for selecting stable and salient 2D laser interest points using Dirichlet process clustering (DP-means). This method is tested on both synthetic and real data from a forest trail and compared with [1]. The result reveals the importance of salient interest point selection in allowing accurate and fast place recognition. Our approach also ensures a low bandwidth representation of visited areas, making it suitable for real-time, multi-agent SLAM applications.

Whilst Micro Aerial Vehicles (MAVs) possess a variety of promising capabilities, their high energy consumption severely limits applications where flight endurance is of high importance. Reducing energy usage is one of the main challenges in advancing aerial robot utility. To address this bottleneck in the development of unmanned aerial vehicle applications, this work proposes an bioinspired mechanical approach and develops an aerial robotic system for greater endurance enabled by low power station-keeping. The aerial robotic system consists of an multirotor MAV and anchoring modules capable of launching multiple tensile anchors to fixed structures in its operating envelope. The resulting tensile perch is capable of providing a mechanically stabilized mode for high accuracy operation in 3D workspace. We explore generalised geometric and static modelling of the stabilisation concept using screw theory. Following the analytical modelling of the integrated robotic system, the tensile anchoring modules employing high pressure gas actuation are designed, prototyped and then integrated to a quadrotor platform. The presented design is validated with experimental tests, demonstrating the stabilization capability even in a windy environment.

Visual markers are useful tools that aid autonomous robots in object recognition and pose measurement; however, conventional markers have two fundamental problems in orientation estimation: (1) inaccurate estimation in the frontal direction and (2) pose ambiguity. We previously developed a visual marker “LentiMark” that solves problem (1). In this paper, we propose a solution to problem (2) by improving LentiMark. For this purpose, we use two wavelike slices that change two-tone patterns according to the angle of visual line. The new marker enables stable and robust pose estimation without pose ambiguity. We solve the two biggest problems of visual markers in the framework of small planar markers. We verify the efficiency of our method through experiments.

Efforts in soft-matter sensing with liquid metal (LM) have largely focused on pressure, strain, and bend sensing by measuring changes in capacitance or resistance of LM circuits embedded in a soft elastomer. Here, we show that LM electronics can be used to combine pressure sensing with proximity detection and remote finger tracking - sensing modalities not previously demonstrated in a soft electronic skin (E-Skin). This is accomplished with a novel architecture composed of 9 capacitive sensors arranged in two layers and placed below a spiral-shaped pressure transducer. The E-skin is composed of eutectic gallium-indium (EGaIn) LM alloy and Pt-cured EcoFlex 0030 elastomer. The multi-modal sensor is able to detect the proximity and motion of a finger as it moves above the sensor and is able to detect touch and pressure once the finger comes into contact. Proximity sensing and contact pressure are distinguish by comparing changes in capacitance and resistance of the embedded LM circuitry. Measurements on a cm-scale prototype suggest proximity sensing with a range of 8.7 cm and maximum contact pressure of 110 kPa.

Precise patient positioning is fundamental to successful removal of malignant tumors during treatment of head and neck cancers. Errors in patient positioning have been known to damage critical organs and cause complications. To better address issues of patient positioning and motion, we introduce a 3-DOF neuro-adaptive soft-robot, called Soft-NeuroAdapt to correct deviations along 3 axes. The robot consists of inflatable air bladders that adaptively control head deviations from target while ensuring patient safety and comfort. The adaptive-neuro controller combines a state feedback component, a feedforward regulator, and a neural network that ensures correct adaptation. States are measured by a 3D vision system. We validate Soft-NeuroAdapt on a 3D printed head-and-neck dummy, and demonstrate that the controller provides adaptive actuation that compensates for intrafractional deviations in patient positioning.

We present a fully edible pneumatic actuator based on gelatin-glycerol material. The actuator is monolithic, fabricated via a molding process, and measures 90 mm in length, 20 mm in width, and 17 mm in thickness. Thanks to the material mechanical characteristics similar to those of silicone elastomers, the actuator exhibits a bending angle of 170.3 ° and a blocked force of 0.34 N at the applied pressure of 25 kPa. These values are comparable to elastomer based pneumatic actuators. As a validation example, two actuators are integrated to form a gripper capable of handling various objects, highlighting the high performance and applicability of the edible actuator. These edible actuators, combined with other recent edible materials and electronics, could lay the foundation for a new type of edible robots.