Event cameras asynchronously report timestamped changes in pixel intensity and offer advantages over conventional raster scan cameras in terms of low-latency, low redundancy sensing and high dynamic range. In recent years, much of research in event based vision has been focused on performing tasks such as optic flow estimation, moving object segmentation, feature tracking, camera rotation estimation and more, through contrast maximization. In contrast maximization, events are warped along motion trajectories whose parameters depend on the quantity being estimated, to some time t_ref. The parameters are then scored by some reward function of the accumulated events at t_ref. The versatility of this approach has lead to a flurry of research in recent years, but no in-depth study of the reward chosen during optimization has yet been made. In this work we examine the choice of reward used in contrast maximization, propose a classification of different rewards and show how a reward can be constructed that is more robust to noise and aperture uncertainty. We validate our work experimentally by predicting optical flow and comparing to ground-truth data.

In contrast to traditional cameras, whose pixels have a common exposure time, event-based cameras are novel bio-inspired sensors whose pixels work independently and asynchronously output intensity changes (called "events"), with microsecond resolution. Since events are caused by the apparent motion of objects, event-based cameras sample visual information based on the scene dynamics and are, therefore, a more natural fit than traditional cameras to acquire motion, especially at high speeds, where traditional cameras suffer from motion blur. However, distinguishing between events caused by different moving objects and by the camera's ego-motion is a challenging task. We present the first per-event segmentation method for splitting a scene into independently moving objects. Our method jointly estimates the event-object associations (i.e., segmentation) and the motion parameters of the objects (or the background) by maximization of an objective function, which builds upon recent results on event-based motion-compensation. We provide a thorough evaluation of our method on a public dataset, outperforming the state-of-the-art by as much as 10%. We also show the first quantitative evaluation of a segmentation algorithm for event cameras, yielding around 90% accuracy at 4 pixels relative displacement.

In this paper, we present the first multilevel implementation of the Harris-Stephens corner detector and the ORB feature extractor running on FPGA hardware, for computer vision and robotics applications. ORB is a fundamental component of many robotics applications, and requires significant computation. The design has been validated both in behavioural simulation and in implementation on an Arria V FPGA connected to a desktop PC via PCI-Express. A Linux kernel-mode driver and userspace library allow integration of the acceleration hardware into C++ programs. The device has significantly higher throughput than a CPU implementation (150 MPixel/s vs 27 MPixel/s) and a GPU implementation (40 MPixel/s), with much lower power draw (5.3 W vs 145 W). This throughput is equivalent to 72 fps at 1920 × 1080 or 488 fps at 640 × 480.

This brief presents a parallel single-rail self-timed adder. It is based on a recursive formulation for performing multibit binary addition. The operation is parallel for those bits that do not need any carry chain propagation. Thus, the design attains logarithmic performance over random operand conditions without any special speedup circuitry or look-ahead schema. A practical implementation is provided along with a completion detection unit. The implementation is regular and does not have any practical limitations of high fanouts. A high fan-in gate is required though but this is unavoidable for asynchronous logic and is managed by connecting the transistors in parallel. Simulations have been performed using an industry standard toolkit that verify the practicality and superiority of the proposed approach over existing asynchronous adders.

Different image processing approaches are presented that are candidates for the Monash Vision Group prosthetic vision device. As described in a companion paper [6], the Monash Vision Group is developing a bionic eye based on the implantation of 7-11 stimulation tiles on the primary visual cortex of the brain. In the lead up to an expected 2014 first in-human trial of this device, potential image processing techniques and intuitive user interfaces need to be developed and evaluated. An FPGA Hatpack has been developed to give normally sighted people a similar binary limited resolution visual experience expected of a Bionic Eye recipient. The Hatpack has been used to generate new psychophysics results to evaluate and improve the performance and practicality of three different methods of luminance threshold selection. The results show an interesting difference in terms of performance and highlight areas of possible improvement for the algorithms used.

Visual prostheses such as retinal implants provide bionic vision that is limited in spatial and intensity resolution. This limitation is a fundamental challenge of bionic vision as it severely truncates salient visual information. We propose to address this challenge by performing real time transformations of visual and non-visual sensor data into symbolic representations that are then rendered as low resolution vision; a concept we call Transformative Reality. For example, a depth camera allows the detection of empty ground in cluttered environments that is then visually rendered as bionic vision to enable indoor navigation. Such symbolic representations are similar to virtual content overlays used in Augmented Reality but are registered to the 3D world via the user's sense of touch. Preliminary user trials, where a head mounted display artificially constrains vision to a 25×25 grid of binary dots, suggest that Transformative Reality provides practical and significant improvements over traditional bionic vision in tasks such as indoor navigation, object localisation and people detection.

This paper describes a probabilistic online map merging system for a single mobile robot. It performs intermittent exploration by fusing laser scan matching and omnidirectional vision. Moreover, it can also be adapted to a multi-robot system for large scale environments. Map merging is achieved by means of a probabilistic Haar-based place recognition system using omnidirectional images and is capable of discriminating new and previously visited locations in the current or previously collected maps. This dramatically reduces the search space for laser scan matching. The combination of laser range finding and omnidirectional vision is very attractive because they reinforce one another when there is sufficient structure and visual information in the environment. In other cases, they complement one another, leading to improved robustness of the system. This is the first system to combine a probabilistic Haar-based place recognition system using omnidirectional images with laser range finding to merge maps. The proposed system is also algorithmically simple, efficient and does not require any offline processing. Experimental results of the approach clearly illustrate that the proposed system can perform both online map merging and exploration robustly using a single robot configuration in a real indoor lab environment.

Advanced sonar systems produce both accurate range and bearing measurements rather than just range alone as in conventional systems. Previous advanced sonar rings do not process incoming echo data as it arrives, but only after a completed measurement cycle and only on limited data samples that are above a noise floor threshold. The system described in this paper can process all echo data, as it is produced, with full matched filtering tuned to each receiver on all echo samples directly implemented in hardware. This minimises measurement latency, important for real time robotics applications and also provides optimal arrival time estimates due to the matched filters. Previous systems have used sequential firing of many transmitters around the ring to prevent interference between transmitters or used different pulses fired simultaneously. This paper presents a single transmitter solution where a reflector disperses a single ultrasonic pulse evenly around the sonar ring. Processing is performed with a dedicated hardware data processing architecture implemented with a Field Programmable Gate Array to achieve the desired real time performance. The system reports range and bearing results at a rate of 30 measurement cycles a second in a full 360 degree coverage to ranges up to 4 metres with the prototype processing over 4.9 Giga-arithmetic operations per second. Experimental results are presented that show the performance of the system is suitable for high speed mapping and localisation applications.

This paper describes a robotic system that learns visual models of symmetric objects autonomously. Our robot learns by physically interacting with an object using its end effector. This departs from eye-in-hand systems that move the camera while keeping the scene static. Our robot leverages a simple nudge action to obtain the motion segmentation of an object in stereo. The robot uses the segmentation results to pick up the object. The robot collects training images by rotating the grasped object in front of a camera. Robotic experiments show that this interactive object learning approach can deal with top-heavy and fragile objects. Trials confirm that the robot-learned object models allow robust object recognition.

SLAM is a well studied technique for robots to build a map of environments while at the same time keeping track of their pose (position and orientation). However SLAM does not provide control approaches for how the robot moves around the environment. This paper presents an integrated approach to create a fully autonomous exploring and mapping robot. An EKF-SLAM approach is used to fuse Advanced Sonar and Laser Scan-Matching. This also tackles the problem of map-drifts in some types of environment where lasers do not supply sufficient information in some directions such as along a corridor. In addition, the proposed exploration algorithm takes advantage of the characteristic of the Voronoi Graph to enable the robot to strategically explore the environments in a loop-closing fashion and safe manner. By revisiting areas to close loops as early as possible, the robot can build a more stable map incrementally while still reliably tracking its pose. Experimental results of the integrated approach are shown to demonstrate the algorithm provides real-time exploration of a mobile robot in an initially unknown real environment. Experimental comparisons of exploration strategies with and without early local loop closing demonstrate the benefits of the approach in the map quality.

Location awareness remains the key for many potential future applications of distributed wireless ad hoc sensor networks (WSNs). While the location of a WSN node can be estimated by incorporating Global Positioning System (GPS) devices, it is not suitable to embed GPS receivers in every node considering the cost and size of these devices as well as from an optimization point of view. However, a small number of WSNs nodes called anchor nodes are able to resolve their location either through fixed deployment or using GPS receivers, and thereby provide the reference framework for localization of other nodes. The measurement devices in individual nodes are often erroneous for tiny WSNs nodes, and hence, robustness is a major issue for localization. In this paper, a theoretical localization framework in the presence of noise is postulated, which achieves accurate positioning compared to the existing theoretical approaches. The paired measurement localization (PML) strategy is evaluated through simulations under various noise conditions and environmental modeling, and practically verified by a testbed implementation with real motes. The results corroborate the improved positioning as well as the robustness of PML for ad hoc wireless sensor networks in the presence of noise.

This paper details a robust and accurate segmentation method for near-symmetric objects placed on a table of known geometry. Here we define visual segmentation as the problem of isolating all portions of an image that belongs to a physically coherent object. The term near-symmetric is used as our method can segment objects with some non-symmetric parts, such as a coffee mug and its handle. Using bilateral symmetry this problem is solved autonomously and robustly through the aid of physical action provided by a robot manipulator. Our proposed approach does not require prior models of target objects and assumes no previously collected background statistics. Instead, our approach relies on a precise robotic nudge to generate the necessary object motion to perform segmentation. Experiments performed on ten objects show that our model-free approach can autonomously and accurately segment a variety of objects. These experiments also indicate that our segmentation approach is not adversely affected when operating in cluttered scenes and can segment multi-coloured and transparent objects in a robust manner.

Geographic routing a.k.a. location aided routing for mobile ad hoc networks is popular for being able to meet the challenging issue of scalability and efficiency in dynamic networks. In the analysis of geographic routing, it is assumed that all the nodes know their location. On the other hand, localization for WSNs is another challenging research area. As localization error can hinder geographic routing performance significantly, it is essential to have a combined evaluation of localization and geographic routing to understand the tradeoffs among different self-localization schemes for location based routing in WSNs. In this paper, several practical localization methods are evaluated for their suitability to geographic routing by extensive simulations.

Many objects found in domestic environments are reflectionally symmetric. In this paper, we present a system that can visually track moving objects by their reflectional symmetry in real time. Apart from the assumption of symmetry, the tracking system does not require any prior object models of the target, such as its colour and shape. The system is robust to shadows and specular reflections. It can also deal with transparent objects. Block motion detection is used in conjunction with symmetry for object tracking. A Kalman filter is used to estimate the object state. Predictions from the Kalman filter is used to improve the efficiency of the symmetry detector. The tracker provides a real time segmentation of an object by searching for motion that is symmetric about the object's mirror line. The tracking system also generates a rotated bounding box, aligned with the object's symmetry line, which can be used as a window for other image processing operations. The final system can track single objects in 640times480 videos at over 40 frames per second using a standard notebook PC

Symmetry is a salient visual feature of many man-made objects. This paper describes research into the detection and segmentation of reflectionally symmetric objects in digital images, without the use of a priori object models. The detection method does not assume uniform object colour or texture, and does not rely on prebuilt models such as 3D geometric primitives. A novel detection algorithm has been developed to find lines of reflectional symmetry in images. This detection algorithm can operate at 10 frames per second on 640 by 480 pixel images. Using the detected symmetry, objects are segmented with a dynamic programming approach. Both algorithms have been extended to accommodate skew symmetry

This paper introduces a method of building topological maps using sequences of images and the approximate string matching algorithm, which is commonly used in DNA sequence alignment applications. Contrary to many existing dense image based topological localisation techniques that operate in known maps, our method builds topological maps in unknown environments. And unlike traditional topological mapping methods that require the robot to explicitly recognise topological path junctions during exploration, our method requires no such explicit "junction detectors". It receives as input only a sequence of unlabeled images. The validity of the approach has been demonstrated in both indoor and outdoor environments. The largest outdoor map created measures 70 by 40 meters with 3 nested loops. The system has shown robustness towards large amounts of sensor aliasing and noise caused by errors in the path following behaviours

This paper presents a novel approach to mapping for mobile robots that exploits user interaction to semiautonomously create a labelled map of the environment. The robot autonomously follows the user and is provided with a verbal commentary on the current location with phrases such as “Robot, we are in the office”. At the same time, a metric feature map is generated using fusion of laser and advanced sonar measurements in a Kalman filter based SLAM framework, which is later used for localization. When mapping is complete, the robot generates an occupancy grid for use in global task planning. The occupancy grid is created using a novel laser scan registration scheme that relies on storing the path of the robot along with associated local SLAM features during mapping, and later recovering the path by matching the associated local features to the final SLAM map. The occupancy grid is segmented into labelled rooms using an algorithm based on watershed segmentation and integration of the verbal commentary. Experimental results demonstrate our mobile robot creating SLAM and segmented occupancy grid maps of rooms along a 70 metre corridor, and then using these maps to navigate between rooms.

This paper presents a novel method for 2D laser scan matching called polar scan matching (PSM). The method belongs to the family of point to point matching approaches. Our method avoids searching for point associations by simply matching points with the same bearing. This association rule enables the construction of an algorithm faster than the iterative closest point (ICP). Firstly the PSM approach is tested with simulated laser scans. Then the accuracy of our matching algorithm is evaluated from real laser scans from known relative positions to establish a ground truth. Furthermore, to demonstrate the practical usability of the new PSM approach, experimental results from a Kalman filter implementation of simultaneous localization and mapping (SLAM) are provided.

Increasing the information content of measurements can ease some of the problems associated with simultaneous localization and mapping (SLAM). We present an approach for combining measurements from a laser range finder with measurements from an advanced sonar array capable of accurate range and bearing measurements and edge, corner and plane classification. In our approach sonar aids laser segmentation, laser aids good sonar point feature selection and laser and sonar measurements of the same object are fused. We also present a novel approach for fitting right angle corners to laser range data, which enables simple error estimation through the minimization of sum of square range residuals. The results are then used for SLAM with a mobile robot.

Sonar rings are widely used for indoor mobile robots. However, it is difficult to perform on-the-fly applications such as map building and localization using a conventional sonar ring due to low speed, accuracy and interference. Digital signal processing (DSP) techniques and interference rejection ideas are applied in this paper to design a new more sophisticated, fast and accurate sonar ring called an advanced sonar ring. The advanced sonar ring consists of 48 ultrasonic transducers, 24 acting as transceivers and 24 acting as receivers, seven DSP echo processor boards, twelve four-channel 12-bit 500 kHz ADCs and low noise variable gain preamplifiers. The sonar ring is able to cover 360 degrees around robot with simultaneously firing of all 24 transmitters. Transmission and echo analysis are performed at repetition rates of about 15 Hz, depending on the environment, for ranges up to six meters. Accurate distance and bearing measurements of objects are performed in the DSP system using matched filtering techniques. The paper presents new transmit coding based on pulse duration to differentiate neighbouring transmitters in the ring. Experimental data show the effectiveness of the proposed system.

This paper addresses the problem of visual servo control for a humanoid robot in an unstructured domestic environment. The important issues in this application are autonomous planning, robustness to camera and kinematic model errors, large pose errors, occlusions and reliable visual tracking. Conventional image-based or position-based visual servoing schemes do not address these issues, which motivated the proposed hybrid position-based scheme exploiting fusion of visual and kinematic measurements. Kinematic measurements provide robustness to visual distractions, and allow servoing to continue when the end-effector leaves the field of view. Visual measurements provide the complementary benefits of accurate pose tracking and online estimation of the hand-eye transformation for kinematic calibration. Furthermore, it is shown that calibration errors in the focal length and baseline can be approximated as an unknown scale of the end-effector, which can be estimated in the tracking filter to overcome camera calibration errors. The improved accuracy and robustness compared to conventional position-based servoing is demonstrated experimentally.

This paper describes a complete vision-based framework that enables a humanoid robot to perform simple manipulations in a domestic environment. Our system emphasizes autonomous operation with minimal a priori knowledge in an unstructured environment, with robustness to visual distractions and calibration errors. For each new task, the robot first acquires a dense 3D image of the scene using our novel stereoscopic light stripe scanner that rejects secondary reflections and cross-talk. A data-driven analysis of the range map identifies and models simple objects using geometric primitives. Objects are reliably tracked through clutter and occlusions by exploiting multimodal cues (colour, texture and edges). Finally, manipulations are performed by controlling the end-effector using a hybrid position-based visual servoing scheme that fuses visual and kinematic measurements and compensates for calibration errors. Two domestic tasks are implemented to evaluate the performance of the framework: identifying and grasping a yellow box without any prior knowledge of the object, and pouring rice from an inter-actively selected cup into a bowl.

An advanced sonar sensor produces accurate range and bearing measurements, classifies targets and rejects interference with one sensing cycle. Two advanced sonar systems are used to simultaneously localise and map an indoor environment using a mobile robot. This paper presents the approach and results from on-the-fly map building using a Kalman filter and a new odometry error model that incorporates variations in effective wheel separation and angle measurements. This model is suited to pneumatic tyre odometry errors where the wheel separation has been found to vary unpredictably with floor surface and path curvature. The paper also presents techniques for detecting sonar feature clutter and selecting strong candidates for ultrasonic landmarks. The paper illustrates that sonar SLAM data association problems are significantly simplified when advanced sonar sensors are employed compared to Polaroid ranging modules.

This paper presents results from a four transducer pulse coded sonar system that performs target localisation in two dimensions and classification into planes, concave corners and convex edges whilst the sensor is in motion. On each sensing cycle two pulses are emitted from separate transmitters, and two receivers collect echoes that are processed using a DSP. The sensor achieves on-the-fly classification by transmitting nearly and simultaneously from the two transmitters. The effects of sensor motion are analysed in the paper and effects on range and bearing estimation can then be compensated using odometry based robot velocity measurements. Results are presented that show classification and angle measurement deviations from a robot moving at speeds up to 1 metre per second.

This paper presents an integrated, low-level approach to removing sensor noise, cross talk, spurious specular reflections, and solving the association problem in a light stripe scanner. Most single-camera scanners rely on the laser brightness exceeding that of the entire image. Our system uses two cameras to measure the stripe and combines the knowledge of the light plane orientation to produce useful validation properties. The key result is the development of a condition relating image plane measurements and camera intrinsic parameters, which allows validation/association to be performed independently of 3D reconstruction. The same equations are used to improve ranging accuracy compared to single-camera systems. We also show how the system may be self-calibrated using measurements of an arbitrary nonplanar target. As validation allows the operation in ambient light, the registered colour and range are captured in the same sensor. An experimental scanner demonstrates the effectiveness of the proposed techniques.

This paper describes a new sonar system that can perform target localisation in two dimensions and classification into planes, concave corners and convex edges with no extra time overhead, that is, the sensor transmits on two transmitters a short time apart, thereby collecting echoes in virtually the same time as a single transmitter of the system. Moreover, the time separation of the transmitted pulses acts to identify the particular sonar system so that interference from other systems can be rejected. The sensor combines two previous sonar research efforts on double pulse coding and classification in a real time DSP-based sensing module that is also smaller than previous sensors. Since the classification is performed with such a short delay between transmitter firings, the sensor can be deployed on moving platforms to achieve on-the-fly mapping. This paper describes the sonar hardware, maximum likelihood estimation classification approach and experimental results.

This paper describes a new highly accurate fast self-contained sonar sensor. Transmission and echo analyses are performed at repetition rates exceeding 27 Hz for ranges up to 5.4 metres. The sensor contains low noise variable gain preamplifiers, two 1 MHz 12 bit ADC receivers and a DSP echo processor. Optimal arrival time estimation is performed by the DSP using matched filtering of echoes with short duration and wide bandwidth. With the sensor mounted on a mobile robot, map results are derived from scanning with the sensor as the robot moves continuously.

Details the design and construction of a full sized robotic humanoid whose body is made entirely from lightweight polyurethane plastic and whose total size and weight are that of a typical human adult. The design of this robot was conceived with research in the conventional human environment in mind, incorporating all aspects of human mobility and sensory capabilities. Plastic components for the robot are easily and quickly manufactured, and are easy and cost effective to replace if damaged or broken. The substantial saving in weight that is possible through the use of plastics significantly affects other aspects of the design including motor and cabling capacities as well as battery life.

A sonar target tracking system is presented that is capable of accurately tracking targets at measurement rates exceeding 10 Hz. Two sonar trackers, each consisting of a transmitter and two receivers, are independently controlled to track sonar targets from bearing and range measurements. Bearing and range are accurately estimated using matched filters on two closely spaced receivers, with accuracy better than 0.1 degrees in still air conditions. Accuracy degrades with increasing air turbulence and temperature gradients, and bearing errors are shown experimentally to have significant autocorrelation at times of the order of seconds. The transmitter identity is coded using time separation of double pulses, thus allowing the two transmitters to operate simultaneously without spurious crosstalk readings. The ability to reliably reject interference is demonstrated experimentally. A simple computationally lean double pulse validation approach is analysed and experimentally tested with robot speeds up to 1 metre per second where Doppler shifting of the double pulse separation is an important factor.

Conventional wall following with an ultrasonic sensor uses only range data to the nearest reflecting point. However, the bearing angle information to the wall is more useful for a wall following motion. In this paper, we propose a simple wall following algorithm, where the robot moves perpendicular to the direction to the nearest reflecting point. Also, in conventional ultrasonic pulse-echo sensing, an accurate target bearing measurement is often regarded as difficult due to the wide directivity of ultrasonic transducers. However, by assuming that the ultrasonic echo returns from a single direction, the bearing angle can be measured. A new sensing method is also proposed to determine accurately the bearing angle to the reflecting point by a single ultrasonic transducer. This paper also presents experimental results from mobile robot wall following experiments using only bearing information measured by a single ultrasonic transducer. These experiments illustrate the effectiveness of proposed method.

This paper describes a mobile robot equipped with a sonar sensor array, Werrimbi, in a guided feature based map building task in an indoor environment. Common indoor landmarks such as planes, corners and edges are located and classified with a multiple transducer sensor array. Accurate odometry information is derived from a pair of narrow unloaded encoder wheels. Discrete sonar observations are incrementally merged into partial planes to produce a realistic representation of environment. Collinearity constraints among features are exploited to enhance state estimation. The map update utilises Julier-Uhlmann Kalman filter which improves the accuracy of covariance propagation through nonlinear equations and eliminates the need to derive Jacobian matrices. Correlation among map features and robot location are explicitly represented. Partial planes are also used to eliminate phantom targets caused by sonar specular reflection.

This paper presents the key steps involved in the design, calibration and error modelling of a low cost odometry system capable of achieving high accuracy dead-reckoning. A consistent error model for estimating position and orientation errors has been developed. Previous work on propagating odometry error covariance relies on incrementally updating the covariance matrix in small time steps. The approach taken here sums the noise theoretically over the entire path length to produce simple closed form expressions, allowing efficient covariance matrix updating after the completion of path segments. Systematic errors due to wheel radius and wheel base measurement were first calibrated with UMBmark test. Experimental results show that, despite its low cost, our system's performance, with regard to dead-reckoning accuracy, is comparable to some of the best reported odometry vehicle.

This paper presents an environmental acquisition strategy for a mobile robot using an advanced sonar sensor to achieve mapping navigation in an a priori unknown, imperfectly structured indoor environment. Most existing feature based strategies rely on unrealistic assumptions about the environment, while their grid based counterparts hinder localisation which leads to rapid degradation of map quality. A dual representation strategy is proposed here which exploits the strength of both a feature map and a grid map. With the advantage sensor, the environment is scanned and the obtained features are classified into planes, corners, edges and unknowns. The feature map is only updated with the first three types of features. The grid map is updated with all measurement, including the unknowns resulting from complicated objects, to enable obstacle avoidance. On the grid map, distance transform based exploratory path planning is implemented. Adaptation has been made so that an explore-local-first behaviour is exhibited.

Describes a sensor that fuses sonar and visual data to create a three dimensional (3D) model of the environment with application to robot navigation. The environment is characterized by a set of connected horizontal and vertical lines. 3D sonar data is augmented by making deductions concerning the connection and definition of lines in 2D visual data. Any errors that may result from incorrect interpretation of the 2D camera data, such as false connections between lines, can be detected by moving the robot. Experimental results from the sensor are presented.

A sonar system is presented that relies on scanning a single ultrasonic transducer and measuring echo amplitude and arrival times. Bearing angles to targets are estimated far more accurately than the transducer beamwidth as obtained with conventional sonar rings based on the Polaroid ranging module. A Gaussian beam characteristic is fitted using least squares to the amplitudes of corresponding echoes in the scan to obtain an estimate of the bearing to specular targets. As an illustration of the information gain over conventional sonar rings, the sensor approach is used on a mobile robot to find, traverse and map doorways reliably and with minimal algorithmic effort. This is compared with other work that claims the problem is difficult to solve using a conventional sonar ring of 24 Polaroid ranging modules.

This paper presents a novel sonar sensor consisting of three transmitters and three receivers that can localise and classify 3D targets into 16 different naturally occurring indoor classes. The sensor produces submillimeter range and sub-degree bearing accuracies using an optimal matched filter time of flight estimator up to a range of 6 meters. The sensor configuration, hardware and processing are described. Experimental results from the sensor are presented.

This paper describes an ultrasonic sensor which uses the times of flight from three Polaroid ultrasonic transducers arranged in an equilateral triangle to identify and localise planes, 2D and 3D corners. The sensor employs a maximum likelihood estimator and a data acquisition system with a low sampling rate of about 59 kHz. The hardware and processing requirements are modest and fast due to the simple identification algorithms and sensor structure. Localisation of the objects can be achieved with range error of about 2 mm and bearing error of less than 1/spl deg/. The sensor has been applied to localising a robot in a known indoor environment using 3D natural features and has achieved accuracies of 1 cm in position and 2/spl deg/ in bearing.

A novel sonar array for mobile robots is presented with applications to localization and mapping of indoor environments. The ultrasonic sensor localizes and classifies multiple targets in two dimensions to ranges of up to 8 meters. By accounting for effects of temperature and humidity, the system is accurate to within 1 mm and 6.1 degrees in still air. Targets separated by 10 mm can be discriminated. Targets are classified into planes, corners, edges and unknown, with the minimum of two transmitters and two receivers. A novel approach is that receivers are closely spaced to minimize the correspondence problem of associating echoes from multiple targets. A set of templates is generated for echoes to allow the optimal arrival time to be estimated, and overlapping echoes and disturbances to be rejected.>

Introduces the concept of short-lived navigational markers (SLNMs) which represent a group of techniques for indicating the path taken by a mobile robot. Information provided by marking the robot trajectory can be of direct assistance in a number of navigation tasks. The short-lived nature of the markers has the advantage that it is not necessary to remove them after use. However, the varying intensity of the SLNMs complicates the process of detecting and following the robot path. Heat is one form of SLNM being investigated and this paper describes a pyroelectric sensor which has been developed to detect thermal paths created by heating the floor with a quartz halogen bulb. Three control strategies for following the thermal trail have been developed and tested. A strategy based on Kalman filtering gives the best performance. Details of the control strategies are presented together with experimental results showing how they performed in practice.

The principle of a simple acoustic sensor system using three Polaroid transducers which can identify a plane, an edge, and a corner is discussed. Vector analysis of reflections involving one, two, and three mirror-like orthogonal surfaces is presented. An algorithm to distinguish a corner from an edge using the triangular three-transceiver system is described, and computer simulated results are given.>

An active beacon localization system that estimates position and heading for a mobile robot is described. An iterated extended Kalman filter was applied to the beacon and dead-reckoning data to estimate optimal values of position and heading, given a model for the localizer and robot motion. The author describes the implementation and experimental results of the localization system. Position and heading angle updates were calculated in real time every 150 ms with a measured standard deviation of path error of 40 mm in a 12 m/sup 2/ workspace.>

A synchronizer timing model, called the jitter model, which has general application to metastable reliability analysis, is proposed and analyzed. The jitter model is applied to show that redundancy cannot improve the metastable reliability of synchronizers, contradicting previous work by A. El-Amawy (see ibid., vol.38, no.5, p.750-3 (1989)). The jitter model extends previous synchronizer input timing models by incorporating the effects of circuit noise. The circuit noise translates into jitter or random time displacement of a previously proposed deterministic aperture mode. The jitter model is supported by simulation, circuit analysis, and experimental work. The results of a SPICE simulation of a CMOS D-type flip-flop are presented. An experimental bistable device is constructed to examine the behavior of synchronizers with noise. Statistical results obtained from the experimental bistable device support the jitter model for metastability. The sensitivity of metastable reliability of redundant synchronizers to modeling assumptions is highlighted.>

Fault-free digital systems can fail as a result of metastable behavior when asynchronous inputs have critical timing combinations. The problem of metastable behavior is generally considered to be unavoidable in digital systems that synchronize asynchronous inputs. This correspondence extends previous results on the unavoidability of metastable behavior. The set of inputs to the digital system is generalized to cover a wide range of possibilities encountered in practical circuits whilst still maintaining the result of unavoidability of metastable behavior. A sufficient condition for the set of inputs to contain a range of inputs that excite metastable behavior is that the set of input functions is connected and contains two input functions which drive the system to different stable states.

A class of arbiters, known as batching arbiters, is introduced and defined. A particularly simple decentralised example of a batching arbiter is described, with motivation given for the batching arbiter model adopted. It is shown that under reasonable assumptions, batching arbiters can be described by a finite state Markov chain. The key steps in the analysis of the arbiter performance are the method of assigning states, evaluation of state transition probabilities and showing that the Markov chain is irreducible. Arbiter performance parameters are defined, such as proportion of time allocated to each requester and mean waiting time for each requester. Apart from results describing the steady state behavior of the arbiter for general system parameters, a number of limiting results are also obtained corresponding to light and heavy request loading.

Abstract A simple technique that improves a system's capability of error detection and correction, and enhances software debugging is described in this note. The technique relies on consistency checking of branching variables during branching decisions and the appropriate selection of codes for the permissible values of variables.

This paper considers the possibility of achieving improvements in the reliability of synchronizing an asynchronous signal, by exploiting redundancy and masking. Redundancy and masking techniques have been applied successfully to mask both permanent and transient hardware faults. However, it is shown in this paper that redundancy and masking techniques are ineffective against synchronization failures which arise because of metastable behavior of synchronizing elements.