Organizational Unit:

Institute for Robotics and Intelligent Machines (IRIM)

Permanent Link

https://hdl.handle.net/1853/70657

Includes Organization(s)

Organizational Unit

Biorobotics and Human Modeling Lab

Organizational Unit

Humanoid Robotics Laboratory

Full item page

Publication Search Results

Now showing 1 - 10 of 92

Pulling Open Novel Doors and Drawers with Equilibrium Point Control

(Georgia Institute of Technology, 2009-12) Jain, Advait ; Kemp, Charles C.

A large variety of doors and drawers can be found within human environments. Humans regularly operate these mechanisms without difficulty, even if they have not previously interacted with a particular door or drawer. In this paper, we empirically demonstrate that equilibrium point control can enable a humanoid robot to pull open a variety of doors and drawers without detailed prior models, and infer their kinematics in the process. Our implementation uses a 7 DoF anthropomorphic arm with series elastic actuators (SEAs) at each joint, a hook as an end effector, and low mechanical impedance. For our control scheme, each SEA applies a gravity compensating torque plus a torque from a simulated, torsional, viscoelastic spring. Each virtual spring has constant stiffness and damping, and a variable equilibrium angle. These equilibrium angles form a joint space equilibrium point (JEP), which has a corresponding Cartesian space equilibrium point (CEP) for the arm's end effector. We present two controllers that generate a CEP at each time step (ca. 100 ms) and use inverse kinematics to command the arm with the corresponding JEP. One controller produces a linear CEP trajectory and the other alters its CEP trajectory based on real-time estimates of the mechanism's kinematics. We also present results from empirical evaluations of their performance (108 trials). In these trials, both controllers were robust with respect to variations in the mechanism, the pose of the base, the stiffness of the arm, and the way the handle was hooked. We also tested the more successful controller with 12 distinct mechanisms. In these tests, it was able to open 11 of the 12 mechanisms in a single trial, and successfully categorized the 11 mechanisms as having a rotary or prismatic joint, and opening to the right or left. Additionally, in the 7 out of 8 trials with rotary joints, the robot accurately estimated the location of the axis of rotation.
Effective robot task learning by focusing on task-relevant objects

(Georgia Institute of Technology, 2009-10) Lee, Kyu Hwa ; Lee, Jinhan ; Thomaz, Andrea L. ; Bobick, Aaron F.

In a robot learning from demonstration framework involving environments with many objects, one of the key problems is to decide which objects are relevant to a given task. In this paper, we analyze this problem and propose a biologically-inspired computational model that enables the robot to focus on the task-relevant objects. To filter out incompatible task models, we compute a task relevance value (TRV) for each object, which shows a human demonstrator's implicit indication of the relevance to the task. By combining an intentional action representation with `motionese', our model exhibits recognition capabilities compatible with the way that humans demonstrate. We evaluate the system on demonstrations from five different human subjects, showing its ability to correctly focus on the appropriate objects in these demonstrations.
RF vision: RFID receive signal strength indicator (RSSI) images for sensor fusion and mobile manipulation

(Georgia Institute of Technology, 2009-10) Deyle, Travis ; Nguyen, Hai ; Reynolds, Matt S. ; Kemp, Charles C.

In this work we present a set of integrated methods that enable an RFID-enabled mobile manipulator to approach and grasp an object to which a self-adhesive passive (battery-free) UHF RFID tag has been affixed. Our primary contribution is a new mode of perception that produces images of the spatial distribution of received signal strength indication (RSSI) for each of the tagged objects in an environment. The intensity of each pixel in the 'RSSI image' is the measured RF signal strength for a particular tag in the corresponding direction. We construct these RSSI images by panning and tilting an RFID reader antenna while measuring the RSSI value at each bearing. Additionally, we present a framework for estimating a tagged object's 3D location using fused ID-specific features derived from an RSSI image, a camera image, and a laser range finder scan. We evaluate these methods using a robot with actuated, long-range RFID antennas and finger-mounted short-range antennas. The robot first scans its environment to discover which tagged objects are within range, creates a user interface, orients toward the user-selected object using RF signal strength, estimates the 3D location of the object using an RSSI image with sensor fusion, approaches and grasps the object, and uses its finger-mounted antennas to confirm that the desired object has been grasped. In our tests, the sensor fusion system with an RSSI image correctly located the requested object in 17 out of 18 trials (94.4%), an 11.1% improvement over the system's performance when not using an RSSI image. The robot correctly oriented to the requested object in 8 out of 9 trials (88.9%), and in 3 out of 3 trials the entire system successfully grasped the object selected by the user.
PPS-Tags: Physical, Perceptual and Semantic Tags for Autonomous Mobile Manipulation

(Georgia Institute of Technology, 2009-10) Nguyen, Hai ; Deyle, Travis ; Reynolds, Matt S. ; Kemp, Charles C.

For many promising application areas, autonomous mobile manipulators do not yet exhibit sufficiently robust performance. We propose the use of tags applied to task-relevant locations in human environments in order to help autonomous mobile manipulators physically interact with the location, perceive the location, and understand the location’s semantics. We call these tags physical, perceptual and semantic tags (PPS-tags). We present three examples of PPS-tags, each of which combines compliant and colorful material with a UHF RFID tag. The RFID tag provides a unique identifier that indexes into a semantic database that holds information such as the following: what actions can be performed at the location, how can these actions be performed, and what state changes should be observed upon task success? We also present performance results for our robot operating on a PPS-tagged light switch, rocker light switch, lamp, drawer, and trash can. We tested the robot performing the available actions from 4 distinct locations with each of these 5 tagged devices. For the light switch, rocker light switch, lamp, and trash can, the robot succeeded in all trials (24/24). The robot failed to open the drawer when starting from an oblique angle, and thus succeeded in 6 out of 8 trials. We also tested the ability of the robot to detect failure in unusual circumstances, such as the lamp being unplugged and the drawer being stuck.
Modeling of biodynamic feedthrough in backhoe operation

(Georgia Institute of Technology, 2009-10) Humphreys, Heather C. ; Book, Wayne J. ; Huggins, James D.

An advanced backhoe user interface has been developed which uses coordinated control with haptic feedback. Results indicate that the coordinated control provides more intuitive operation that is easy to learn, and the haptic feedback also relays meaningful information back to the user in the form of force signals from digging forces and system limitations. However, results show that the current system has significant problems with biodynamic feedthrough, where the motion of the controlled device excites motion of the operator, resulting in undesirable forces applied to the input device and control performance degradation. This unwanted input is difficult to decouple from the intentional operator input in experiments. This research presents an investigation on the effects of biodynamic feedthrough on this particular backhoe control system, using system identification to empirically define models to represent each component. These models are used for a preliminary simulation study on potential methods for biodynamic feedthrough compensation.
GroupSAC: Efficient Consensus in the Presence of Groupings

(Georgia Institute of Technology, 2009-09) Ni, Kai ; Jin, Hailin ; Dellaert, Frank

We present a novel variant of the RANSAC algorithm that is much more efficient, in particular when dealing with problems with low inlier ratios. Our algorithm assumes that there exists some grouping in the data, based on which we introduce a new binomial mixture model rather than the simple binomial model as used in RANSAC. We prove that in the new model it is more efficient to sample data from a smaller numbers of groups and groups with more tentative correspondences, which leads to a new sampling procedure that uses progressive numbers of groups. We demonstrate our algorithm on two classical geometric vision problems: wide-baseline matching and camera resectioning. The experiments show that the algorithm serves as a general framework that works well with three possible grouping strategies investigated in this paper, including a novel optical flow based clustering approach. The results show that our algorithm is able to achieve a significant performance gain compared to the standard RANSAC and PROSAC.
Kinematics and Verification of a Deboning Device

(Georgia Institute of Technology, 2009-08) Zhou, Debao ; Daley, Wayne ; McMurray, Gary

Poultry deboning process is one of the largest employers in the United States and mainly involves human workers due to the unstructured nature of the task. For the automation of this process, a cutting device with the adaptive capability has been developed. In this paper, we focused on the kinematics of this device and the accuracy of the actual cutting point location. We validated the kinematic formulation and proofed the confidence of the accurate cutting. The applied verification method can be generalized to be applicable to general kinematics verification.
A novel approach to fabric control for automated sewing

(Georgia Institute of Technology, 2009-07) Winck, Ryder C. ; Dickerson, Stephen L. ; Huggins, James D. ; Book, Wayne J.

This paper describes a novel fabric manipulation method for fabric control during the sewing process. It addresses issues with past attempts concerning fabric position and tension control. The method described involves replacing the current sewing feed mechanism with a servo controlled manipulator to both feed and control the fabric. The manipulator is coupled with a machine vision system that tracks the threads of the fabric to provide real-time position control that is robust with respect to fabric deformations. A prototype of the manipulator is used to demonstrate the feasibility of the concept, reaching accelerations up to 27 g’s and following a closed loop trajectory with open loop control while operating in coordination with an industrial sewing machine. The system described also offers a general solution to high accuracy and high acceleration position control systems.
Behavior-Based Door Opening with Equilibrium Point Control

(Georgia Institute of Technology, 2009-06) Jain, Advait ; Kemp, Charles C.

Within this paper we present a set of behaviors that enable a mobile manipulator to reliably open a variety of doors. After a user designates a location within 20cm of the door handle, the robot autonomously locates the door handle using a tilting laser range finder, approaches the handle using its omnidirectional base, reaches out to haptically find the door, makes contact with the handle, twists it, and pushes open the door. The robot uses equilibrium point control for all arm motions. Our implementation uses a 7 DoF anthropomorphic arm with series elastic actuators (SEAs). For our control scheme, each SEA applies a gravity compensating torque plus a torque from a simulated, torsional, viscoelastic spring. Each virtual spring has constant stiffness and damping, and a variable equilibrium point. The behaviors use inverse kinematics to generate trajectories for these joint-space equilibrium points that correspond with Cartesian equilibrium point trajectories for the end effector. With 43 trials and 8 different doors, we show that these compliant trajectories enable the robot to robustly reach out to make contact with doors (100%), operate door handles (96.9%), and push doors open (100%). The complete system including perception and navigation succeeded with unlocked doors in 28 out of 32 trials (87.5%) and locked doors in 8 out of 8 trials (100%). Through 157 trials with a single door, we empirically show that our method for door handle twisting reduces interaction forces and is robust to variations in arm stiffness, the end effector trajectory, and the friction between the end effector and the handle.
Bayesian Surprise and Landmark Detection

(Georgia Institute of Technology, 2009-05) Ranganathan, Ananth ; Dellaert, Frank

Automatic detection of landmarks, usually special places in the environment such as gateways, for topological mapping has proven to be a difficult task. We present the use of Bayesian surprise, introduced in computer vision, for landmark detection. Further, we provide a novel hierarchical, graphical model for the appearance of a place and use this model to perform surprise-based landmark detection. Our scheme is agnostic to the sensor type, and we demonstrate this by implementing a simple laser model for computing surprise. We evaluate our landmark detector using appearance and laser measurements in the context of a topological mapping algorithm, thus demonstrating the practical applicability of the detector.