Model drift is a challenging issue for visual object tracking. Most available approaches aim to address this issue by constructing of a stronger discriminative prediction model with the integration of multiple features. In this article, we also address the issue of model drift using multiple features, but we consider the fusion problem from a different point of view, namely, strengthening the features that are suitable for the current scenario while weakening the remaining ones. Therefore, an advanced tracker that adaptively redetermines the reliability for each feature during the tracking process is proposed. Furthermore, to correctly evaluate and redetermine these reliabilities, two different solutions, called model evaluation and numerical optimization, are proposed, and two independent trackers corresponding to these two solutions are implemented. Extensive experiments have been designed on five large datasets to validate the following: 1) The proposed tracking framework is superior for making the tracking model more robust, and 2) the two solutions proposed for redetermining the reliability for each feature are effective. As expected, the two implemented trackers do indeed improve the accuracy and robustness compared to state-of-the-art trackers. Especially on VOT2016, the proposed trackers based on model evaluation and numerical optimization achieve outstanding EAO scores (i.e.,0.453and0.428, respectively), outperforming the recently developed top trackers by a large margin.

By leveraging large amounts of training data and deep learning technologies, performances of modern handwritten optical character recognition (OCR) systems have been greatly improved. However, collecting and labeling massive handwriting images are both time-consuming and expensive. In this paper, we propose to augment handwritten OCR training with online handwriting samples. To achieve this goal, we propose a style-conditioned generative adversarial network (SC-GAN) with a novel training data pair generation strategy. Then this network is used to transfer the styles of real handwriting images to skeleton images extracted from online handwriting samples to generate photo-realistic text line images. Experimental results on a large scale handwritten OCR task show that the recognition accuracy of our handwritten OCR system is improved by using the augmented synthetic training data.

We equip an ultra-wideband (UWB) node and a 2D LiDAR sensor a.k.a. 2D laser rangefinder on a mobile robot, and place UWB beacon nodes at unknown locations in an unknown environment. All UWB nodes can do ranging with each other thus forming a cooperative sensor network. We propose to fuse the peer-to-peer ranges measured between UWB nodes and laser scanning information, i.e., range measured between robot and nearby objects/obstacles, for simultaneous localization of the robot, all UWB beacons and LiDAR mapping. The fusion is inspired by two facts: 1) LiDAR may improve UWB-only localization accuracy as it gives a more precise and comprehensive picture of the surrounding environment; 2) on the other hand, UWB ranging measurements may remove the error accumulated in the LiDAR-based SLAM algorithm. Our experiments demonstrate that UWB/LiDAR fusion enables drift-free SLAM in real-time based on ranging measurements only.

In this paper, we present a new approach, named as dynamic window approach with collision suppression cone (DWA-CSC), to handle the issue of a robot avoiding moving obstacles in a partially known dynamic environment. The concept of collision suppression cone is firstly introduced to define a probable collision area. When moving obstacles approach this area, the proposed DWA-CSC will be activated to allow the robot avoiding the obstacles smoothly and thus preventing collision with them by suppressing its motion direction and velocity, just like the way of obstacle avoidance done by human beings. Various simulation results show that the proposed DWA-CSC approach indeed achieve the avoidance of moving obstacles safely and smoothly.

In this paper, we propose a discriminative and robust appearance model based on features extracted from a random partition image hashing algorithm to account for severe occlusion and disappearance. We divide the original image into multiple sub-blocks with random positions and scales. Hash functions are used to map blocks into compact binary codes, with which more effective target matching can be achieved. The tracking task is then formulated by producing a confidence map for the target and background, and obtaining the best samples using maximum a posteriori estimate. Experimental results demonstrate that our tracker can achieve more accurate tracking results in situations of occlusion, out-of-view, and violent motion blur when compared with most of state-of-the-art competing algorithms. Besides, the proposed tracking algorithm is able to run in real time.

The paper is concerned with the multi-robot leader-following problem in the presence of frequent dropouts in vision detection. In many scenarios, for instance a structured environment, it is inevitable to experience outage of vision detection due to reasons such as the target moving out of view, vision occlusion, motion blurring, etc. The paper proposes a Bayesian trajectory estimation based leader-following approach that can offer accurate path following given intermittent vision observations. The follower robot estimates the trajectory of the leader robot based on the noise-corrupted odometry information of both robots, and inter-robot relative observations based on detection of fiducial markers using an RGBD camera. A linear trajectory-following control method is employed to track a historical pose of the leader robot on the estimated trajectory. Results are obtained based on evaluating the proposed leader-following approach in tests with a zig-zag shaped trajectory and with a trajectory that contains sharp turns.

In this paper, we deal with the problem of image-based visual tracking for mobile robots. It is noted that the presence of actuator dynamics and the unknown target motion increases the complexity of the system model and makes the design of the controller more difficult. To solve this problem, we propose an adaptive control approach via utilizing backstepping technique and extended state observer (ESO). For the controller design, the adaptive technique is employed to estimate the bound of the target motion and the adaptive law is derived from the Lyapunov stability theory. We also adopt two ESOs to estimate and compensate for the disturbances affecting the mobile robot dynamics and the wheel actuator dynamics. It is shown that the proposed adaptive controller guarantees the boundness of all the signals in the closed-loop system and enables the tracking errors to exponentially converge to a compact set which is adjustable.