Robot Builders

Recent research concerning the Gaussian canonical form for Simultaneous Localization and Mapping (SLAM) has given rise to a handful of algorithms that attempt to solve the SLAM scalability problem for arbitrarily large environments. One such estimator that has received due attention is the Sparse Extended Information Filter (SEIF) proposed by Thrun et al., which is reported to be nearly constant time, irrespective of the size of the map. The key to the SEIF’s scalability is to prune weak links in what is a dense information (inverse covariance) matrix to achieve a sparse approximation that allows for efficient, scalable SLAM. We demonstrate that the SEIF sparsification strategy yields error estimates that are overconfident when expressed in the global reference frame, while empirical results show that relative map consistency is maintained.

In this paper, we propose an alternative scalable estimator based on an information form that maintains sparsity while preserving consistency. The paper describes a method for controlling the population of the information matrix, whereby we track a modified version of the SLAM posterior, essentially by ignoring a small fraction of temporal measurements. In this manner, the Exactly Sparse Extended Information Filter (ESEIF) performs inference over a model that is conservative relative to the standard Gaussian distribution. We compare our algorithm to the SEIF and standard EKF both in simulation as well as on two nonlinear datasets. The results convincingly show that our method yields conservative estimates for the robot pose and map that are nearly identical to those of the EKF.

read story

Admit it: when you look at a spreadsheet, you trust the results. You don’t question the totals or the percentages you see. But data, and formulas used to compute what appear in the spreadsheet cells, are both prone to errors. According to computer scientists from Oregon State University (OSU), 90% of the 100 million spreadsheets produced each year in the U.S. alone, contain non-trivial errors. This is why they’ve developed a semi-automatic debugger for spreadsheet systems. They say that this new approach could save billions of dollars annually. As they’ve licensed their software to a new company in Oregon, this technology might be one day included in the spreadsheet tools we use. read story

The paper reports studies on the motion planning problem for planar star-shaped manipulators. These manipulators are formed by joining k “legs” to a common point (like the thorax of an insect) and then fixing the “feet” to the ground. The result is a planar parallel manipulator with k - 1 independent closed loops. A topological analysis is used to understand the global structure of the configuration space so that the planning problem can be solved exactly. The worst-case complexity of the algorithm is O(k³ N ³), where N is the maximum number of links in a leg. Examples illustrating the method are given.

As robots get smaller, smarter and cheaper they slowly but surely invade our homes. And while the smart home is certainly not a new concept (compare the Xanadu houses built in 1979), some recent projects are moving beyond extensions to the clapper switch: Jason Johnson at the University of Virginia School of Architecture has started a new program called “Robotic Ecologies”. As outlined in a brief article, Johnson is working on buildings that reconfigure their shape according to external conditions like wind and sun as well as to the movements of people inside - check out his blog. Frederic Kaplan at the EPFL is working on active furniture for a new learning center to open end of 2008. And finally Philips is following a similar trend with their HomeLab, integrating their home entertainment systems with interactive toys and mirrors. As Roland Piquepaille notes on Slashdot: “Maybe one day, we’ll talk to our homes and they’ll answer.” And with the possibility of a rapid prototyping robot house builder still open, maybe they’ll do far more than that… -more-

The main contribution of this paper is the reformulation of the simultaneous localization and mapping (SLAM) problem for mobile robots such that the mapping and localization can be treated as two concurrent yet separated processes: D-SLAM (decoupled SLAM). It is shown that SLAM with a range and bearing sensor in an environment populated with point features can be decoupled into solving a nonlinear static estimation problem for mapping and a low-dimensional dynamic estimation problem for localization. This is achieved by transforming the measurement vector into two parts: one containing information relating features in the map and another with information relating the map and robot. It is shown that the new formulation results in an exactly sparse information matrix for mapping when it is solved using an Extended Information Filter (EIF).Thus a significant saving in the computational effort can be achieved for large-scale problems by exploiting the special properties of sparse matrices. An important feature of D-SLAM is that the correlation among features in the map are still kept and it is demonstrated that the uncertainty of the feature estimates monotonically decreases. The algorithm is illustrated and evaluated through computer simulations and experiments.

As robots get smaller, smarter and cheaper they slowly but surely invade our homes. And while the smart home is certainly not a new concept (compare the Xanadu houses built in 1979), some recent projects are moving beyond extensions to the clapper switch: Jason Johnson at the University of Virginia School of Architecture has started a new program called “Robotic Ecologies”. As outlined in a brief article, Johnson is working on buildings that reconfigure their shape according to external conditions like wind and sun as well as to the movements of people inside - check out his blog. Frederic Kaplan at the EPFL is working on active furniture for a new learning center to open end of 2008. And finally Philips is following a similar trend with their HomeLab, integrating their home entertainment systems with interactive toys and mirrors. As Roland Piquepaille notes on Slashdot: “Maybe one day, we’ll talk to our homes and they’ll answer.” And with the possibility of a rapid prototyping robot house builder still open, maybe they’ll do far more than that… -more-

A neurocontroller is described which generates the basic locomotion and controls the sensor-driven behavior of a four-legged and a six-legged walking machine. The controller utilizes discrete-time neurodynamics, and is of modular structure. One module is for processing sensor signals, one is a neural oscillator network serving as a central pattern generator, and the third one is a so-called velocity regulating network. These modules are small and their structures and their functionalities are analyzable. In combination, they enable the machines to autonomously explore an unknown environment, to avoid obstacles, and to escape from corners or deadlock situations. The neurocontroller was developed and tested first using a physical simulation environment, and then it was successfully transferred to the physical walking machines. Locomotion is based on a gait where the diagonal legs are paired and move together, e.g. trot gait for the four-legged walking machine and tripod gait for the six-legged walking machine. The controller developed is universal in the sense that it can easily be adapted to different types of even-legged walking machines without changing the internal structure and its parameters.

The University of Virginia (UVA) School of Architecture has started a new program about ‘robotic ecologies’ which wants to answer the question: Will robots take over architecture? As said the program leader, “This research is not just about architectural machines that move. It is about groups of architectural machines that move with intelligence.” Apparently, buildings tracking our movements and adapting their shape or texture according human presence are not far fetched. Maybe one day, we’ll talk to our homes and they’ll answer... read story

Posture sensing techniques for Compliant Framed Modular Mobile Robots (CFMMR) are presented in this paper using a new Relative Posture Sensor (RPS) combined with standard sensors in a tiered fusion algorithm. The RPS consists of a compliant frame member instrumented with strain gauges and associated algorithms such that the RPS can predict relative posture. The first tier of the fusion algorithm uses traditional Kalman filters and rigid axle kinematic models to predict the global posture of each axle. In the second tier, a Relative Measurement Stochastic Posture Error Correction (RMSPEC) algorithm is introduced to fuse disparate axle data using the RPS. Experimental results are derived from over 60 trials operating the robot on high traction carpet, low traction sand, and sand with rugged rocky terrain. Results comparing the proposed sensory system with standard sensory systems demonstrate that the proposed techniques yield accurate relative posture estimates and posture regulation even on rugged terrain, which is a vast improvement over previous results.