Compared with conventional image sensors, event cameras have been attracting attention thanks to their potential in environments under fast motion and high dynamic range (HDR). To tackle the lost-track issue due illumination changes HDR scene such as tunnels, an object tracking framework has presented based on count images from camera. The contains offline-trained detector online-trained tracker which complement each other: benefits pre-labelled data during training, but may false or missing...

In this study, the authors present role playing learning scheme for a mobile robot to navigate socially with its human companion in populated environments. Neural networks (NNs) are constructed parameterise stochastic policy that directly maps sensory data collected by velocity outputs, while respecting set of social norms. An efficient simulative environment is built and pedestrians trajectories from number real-world crowd sets. each iteration, equipped NN created virtually play itself as...

The intrinsic reaction pathway for the hydrogenation of dimethyl succinate to γ-butyrolactone using CuZnAl catalyst is proposed, and comparison between experimental values fitted highlights accuracy model.

In the initial alignment process of strapdown inertial navigation system (SINS), large misalignment angles always bring nonlinear problem, which can usually be processed using scaled unscented Kalman filter (SUKF). this paper, problem in SINS is further investigated, and strong tracking (STSUKF) proposed with fixed parameters to improve convergence speed, while these are artificially constructed uncertain real application. To stability reduce selection, paper proposes a fuzzy adaptive...

There are two main challenges, drift and scale ambiguity, restricting monocular visual odometry from an extensive application on real autonomous navigation. In this paper, iterative localization framework is presented to globally localize a mobile vehicle equipped with single camera freely available digital map. Inspired by the concept of cloud, new Gaussian-Gaussian Cloud model proposed give unified representation measurement randomness ambiguity in odometry. model, collection cloud drops...

Drift and scale ambiguity are two main issues which reduce localization accuracy in monocular visual odometry (MVO). It is necessary to propose a unified model represent these measurement uncertainties. In this paper, we present geometric map-assisted approach for mobile robots equipped with MVO. We the of MVO as group particles, obey uniform-Gaussian distribution cover uncertainties including randomness. The saliency each particle can be obtained from indicate raw certainty Geometric shape...

In the dual-axis rotation inertial navigation system (INS), besides gyro error, accelerometer rolling misalignment angle and gimbal shaft swing axis non-orthogonal also affect attitude accuracy. Through analysis of structure, we can see that will produce coning errors which cause fluctuation attitude. According to vector, it be seen error generate additional drift velocity along rotating shaft, reduce precision system. this paper, based on establishment modulation average frame, vector...

This study presents a new perspective for autonomous mobile robots path searching by proposing biasing direction towards causal entropy maximisation during random tree generation. Maximum entropy-biased rapidly-exploring (ME-RRT) is proposed where the computed from sampling and integral approximation, incorporated into existing (RRT) planner. Properties of ME-RRT including degenerating conditions additional time complexity are also discussed. The performance approach studied, results...

This paper proposes a state estimation approach 'robust strong tracking unscented Kalman filter with unknown inputs' that can be applied to non-linear systems inputs. Specifically, the and measurement equations are linearised by statistical linearisation. Then, equation of input is derived based on weighted least squares method. The multiple suboptimal fading factor introduced into priori error covariance matrix improve ability for inaccuracy system model abrupt change variables caused...

In this paper, heading reference-assisted pose estimation (HRPE) has been proposed to compensate inherent drift of visual odometry (VO) on ground vehicles, where an error is prone grow while the vehicle making turns or in environments with poor features. By introducing a particular orientation as “heading reference,” framework presented incorporate measurements from reference sensors into VO. A graph formulation then represent problem under commonly used optimization model. Simulations and...

Pose estimation is a critical problem in autonomous vehicle navigation, especially circumstances where sensor failure or attacks exist. In this paper, filter-based secure dynamic pose approach proposed such that the can be resilient under possible attacks. Our estimator coincides with conventional Kalman filter when all sensors on vehicles are benign. If less than half of measurement states compromised by randomly occurring deception attacks, it still gives stable estimates states, i.e., an...

Vehicles in urban areas are facing the problem of lacking Global Positioning System (GPS) signals canyon environments. In this article, we present a finite-horizon Unscented Rauch-Tung-Striebel Smoother (URTSS)-based position estimation method for vehicle localization, which uses information from past, present, and near-future. To estimate pose, nonlinear constant velocity state-space model is established. Sensors can only obtain immediate information, describe near-future obtained via...

Pose estimation with state or measurement constraints has been frequent in autonomous vehicle navigation. Aiming at incorporating inherently, this article proposes a dynamic potential field (DPF)-based formulation to represent states, measurements, and on connected Riemannian manifolds. The equation the output are derived DPF forms, which imply probabilistic inference states measurements. Constraints incorporated by projecting points toward constraint subset space space, where representing...

The data from event cameras not only portray contours of moving objects but also contain motion information inherently. Herein, can be used in event-based and frame-based object trackers to ease the challenges occluded association, respectively. In tracker, events within a short interval are accumulated. Within interval, histogram local time measurements (or ‘motion histogram’) is proposed as feature describe target candidate regions. Then mean-shift tracking approach by shifting tracker...

Accurately extracting subjective contents of speech signals and applying it on controlling robots remain to this day a challenging task as well an insistent demand in human-robot interaction (HRI). A simple classification human's intentions may limit the development robots' natural reactions users. Additionally, there should be control system that can understand translate into inputs. This paper proposes intelligent for HRI. The objective is human' s commands via recognizing, quantifying...

Forest fruit harvesting is carried out through two ways: manual and machine harvesting. Mechanical vibration will become the future development trend because of high efficiency short picking cycle this method. However, appropriate frequency, amplitude, action time are required to improve rate. At same time, it ensured that type forest in process minimize damage trees. The frequency amplitude existing relatively single, result energy transfer loss extensive tree damage. Therefore, study...