This paper presents a new method for improving the accuracy of inertial measurement units (IMUs) mounted on land vehicles. The algorithm exploits nonholonomic constraints that govern motion vehicle surface to obtain velocity observation measurements which aid in estimation alignment IMU as well forward vehicle. It is shown this can be achieved without any external sensing provided certain observability conditions are met. A theoretical analysis together with comparison experimental results...

In this paper, we present a novel method to fuse observations from an inertial measurement unit (IMU) and visual sensors, such that initial conditions of the integration, including gravity estimation, can be recovered quickly in linear manner, thus removing any need for special initialization procedures. The algorithm is implemented using graphical simultaneous localization mapping like approach guarantees constant time output. This paper discusses technical aspects work, observability...

This paper describes the development and implementation of a high integrity navigation system, based on combined use Global Positioning System (GPS) an inertial measurement unit (IMU), for autonomous land vehicle applications. The focuses issue achieving required loop in systems. highlights detection possible faults both before during fusion process order to enhance loop. this fault methodology considers low frequency IMU caused by bias sensor readings misalignment unit, from GPS receiver...

An efficient and analytical continuous-curvature path-smoothing algorithm, which fits an ordered sequence of waypoints generated by obstacle-avoidance path planner, is proposed. The algorithm based upon parametric cubic Bézier curves; thus, it inherently closed-form in its expression, the only requires maximum curvature to be defined. is, computational easy implement. Results show effectiveness generating a path, satisfies upper bound-curvature constraint, that less control effort track...

The development of low-cost unmanned aerial vehicles (UAVs) and light weight imaging sensors has resulted in significant interest their use for remote sensing applications. While attention been paid to the collection, calibration, registration mosaicking data collected from small UAVs, interpretation these into semantically meaningful information can still be a laborious task. A standard collection classification work-flow requires manual effort segment size tuning, feature selection...

An unmanned aerial vehicle (UAV) is tasked to explore an unknown environment and map the features it finds, but must do so without use of infrastructure-based localisation systems such as GPS, or any a priori terrain data. The UAV navigates using statistical estimation technique known simultaneous mapping (SLAM) which allows for location well sees. SLAM offers unique approach with potential applications including planetary exploration, when GPS denied (for example under intentional jamming,...

This paper presents a 3D path planing algorithm for an unmanned aerial vehicle (UAV) operating in cluttered natural environments. The satisfies the upper bounded curvature constraint and continuous requirement. In this work greater attention is placed on computational complexity comparison with other path-planning considerations. rapidly-exploring random trees (RRTs) used generation of collision free waypoints. unnecessary waypoints are removed by simple pruning generating piecewise linear...

This paper describes a vision-based obstacle detection and navigation system for use as part of robotic solution the sustainable intensification broad-acre agriculture. To be cost-effective, robotics must competitive with current human-driven farm machinery. Significant costs are in high-end localization sensors. Our demonstrates combination an inexpensive global positioning inertial vision single stereo detection. The design robot, including detailed descriptions three key parts system:...

A new framework which adopts a rapidly-exploring random tree (RRT) path planner with Gaussian process (GP) occupancy map is developed for the navigation and exploration of an unknown but cluttered environment. The GP outputs probability given any selected query point in continuous space thus makes it possible to explore full when used conjunction planner. Furthermore, map-generated embedded collision along lends itself obstacle avoidance. Finally, map-building algorithm extended include...

Understanding and predicting the intention of pedestrians is essential to enable autonomous vehicles mobile robots navigate crowds. This problem becomes increasingly complex when we consider uncertainty multimodality pedestrian motion, as well implicit interactions between members a crowd, including any response vehicle. Our approach, Probabilistic Crowd GAN, extends recent work in trajectory prediction, combining Recurrent Neural Networks (RNNs) with Mixture Density (MDNs) output...

We address the issue of autonomous navigation, that is, ability for a navigation system to provide information about states vehicle without need priori infrastructure such as GPS, beacons, or map. The algorithm is known simultaneous localisation and mapping (SLAM) it terrain aided (TANS) which has capability online map building, simultaneously utilising generated bound errors in solution. Since does not require any initial knowledge location, presents powerful augmentation more importantly,...

This paper presents results of the application simultaneous localisation and map building (SLAM) for an uninhabited aerial vehicle (UAV). Single vision camera inertial measurement unit (IMU) are installed in a UAV platform. The data taken from flight test is used to run SLAM algorithm. Results show that both uncertainty corrected even though model system observation highly non-linear. results, however, also indicate further work observability relationship between drift number location...

Abstract This paper presents the on‐going design and implementation of a robust inertial sensor based simultaneous localization mapping (SLAM) algorithm for an unmanned aerial vehicle (UAV) using bearing‐only observations. A single color vision camera is used to observe terrain from which image points corresponding features in environment are extracted. The SLAM estimates complete six degrees‐of‐freedom motion UAV along with three‐dimensional position environment. An extended Kalman filter...

This paper presents the benefits of using a low cost inertial measurement unit to aid in an implementation inverse depth initialized SLAM hand-held monocular camera. Results are presented with and without observations for different assumed initial ranges features on same dataset. When only camera, scale scene is not observable. As expected, map depends prior used initialize may drift when exploring new terrain, precluding loop closure. The results show that help improve estimated trajectory...

This paper presents a multi-class image segmentation approach to automate fruit segmentation. A feature learning algorithm combined with conditional random field is applied multi-spectral data. Current classification methods used in agriculture scenarios tend use hand crafted application-based features. In contrast, our uses unsupervised automatically capture most relevant features from the property makes robust against variance canopy trees and therefore has potential be different domains....