Online Learning-based Robust Visual Tracking for Autonomous Landing of Unmanned Aerial Vehicles

Autonomous landing is a challenging and important technology for both military and civilian applications of Unmanned Aerial Vehicles (UAVs). In this paper, we present a novel online adaptive visual tracking algorithm for UAVs to land on an arbitrary field (that can be used as the helipad) autonomously at real-time frame rates of more than twenty frames per second. The integration of low-dimensional subspace representation method, online incremental learning approach and hierarchical tracking strategy allows the autolanding task to overcome the problems generated by the challenging situations such as significant appearance change, variant surrounding illumination, partial helipad occlusion, rapid pose variation, onboard mechanical vibration (no video stabilization), low computational capacity and delayed information communication between UAV and Ground Control Station (GCS). The tracking performance of this presented algorithm is evaluated with aerial images from real autolanding flights using manually- labelled ground truth database. The evaluation results show that this new algorithm is highly robust to track the helipad and accurate enough for closing the vision-based control loop.

Real-Time Adaptive Multi-Classifier Multi-Resolution Visual Tracking Framework for Unmanned Aerial Vehicles

This paper presents a novel robust visual tracking framework, based on discriminative method, for Unmanned Aerial Vehicles (UAVs) to track an arbitrary 2D/3D target at real-time frame rates, that is called the Adaptive Multi-Classifier Multi-Resolution (AMCMR) framework. In this framework, adaptive Multiple Classifiers (MC) are updated in the (k-1)th frame-based Multiple Resolutions (MR) structure with compressed positive and negative samples, and then applied them in the kth frame-based Multiple Resolutions (MR) structure to detect the current target. The sample importance has been integrated into this framework to improve the tracking stability and accuracy. The performance of this framework was evaluated with the Ground Truth (GT) in different types of public image databases and real flight-based aerial image datasets firstly, then the framework has been applied in the UAV to inspect the Offshore Floating Platform (OFP). The evaluation and application results show that this framework is more robust, efficient and accurate against the existing state-of-art trackers, overcoming the problems generated by the challenging situations such as obvious appearance change, variant illumination, partial/full target occlusion, blur motion, rapid pose variation and onboard mechanical vibration, among others. To our best knowledge, this is the first work to present this framework for solving the online learning and tracking freewill 2D/3D target problems, and applied it in the UAVs.

Using the Cross-Entropy Method for Control Optimization: A Case Study of See-and-Avoid on Unmanned Aerial Vehicles

This paper presents an adaptation of the Cross-Entropy (CE) method to optimize fuzzy logic controllers. The CE is a recently developed optimization method based on a general Monte-Carlo approach to combinatorial and continuous multi-extremal optimization and importance sampling. This work shows the application of this optimization method to optimize the inputs gains, the location and size of the different membership functions' sets of each variable, as well as the weight of each rule from the rule's base of a fuzzy logic controller (FLC). The control system approach presented in this work was designed to command the orientation of an unmanned aerial vehicle (UAV) to modify its trajectory for avoiding collisions. An onboard looking forward camera was used to sense the environment of the UAV. The information extracted by the image processing algorithm is the only input of the fuzzy control approach to avoid the collision with a predefined object. Real tests with a quadrotor have been done to corroborate the improved behavior of the optimized controllers at different stages of the optimization process.

Heat shock fusion proteins as vehicles for antigen delivery into the major histocompatibility complex class I presentation pathway

Mice immunized with heat shock proteins (hsps) isolated from mouse tumor cells (donor cells) produce CD8 cytotoxic T lymphocytes (CTL) that recognize donor cell peptides in association with the major histocompatibility complex (MHC) class I proteins of the responding mouse. The CTL are induced apparently because peptides noncovalently associated with the isolated hsp molecules can enter the MHC class I antigen processing pathway of professional antigen-presenting cells. Using a recombinant heat shock fusion protein with a large fragment of ovalbumin covalently linked to mycobacterial hsp70, we show here that when the soluble fusion protein was injected without adjuvant into H-2b mice, CTL were produced that recognized an ovalbumin-derived peptide, SIINFEKL, in association with Kb. The peptide is known to arise from natural processing of ovalbumin in H-2b mouse cells, and CTL from the ovalbumin-hsp70-immunized mice and a highly effective CTL clone (4G3) raised against ovalbumin-expressing EL4 tumor cells (EG7-OVA) were equally effective in terms of the concentration of SIINFEKL required for half-maximal lysis in a CTL assay. The mice were also protected against lethal challenge with ovalbumin-expressing melanoma tumor cells. Because large protein fragments or whole proteins serving as fusion partners can be cleaved into short peptides in the MHC class I processing pathway, hsp fusion proteins of the type described here are promising candidates for vaccines aimed at eliciting CD8 CTL in populations of MHC-disparate individuals.

PCDD/F Emissions from Light-Duty Diesel Vehicles Operated under Highway Conditions and a Diesel-Engine Based Power Generator

PCDD/F emissions from three light-duty diesel vehicles–two vans and a passenger car–have been measured in on-road conditions. We propose a new methodology for small vehicles: a sample of exhaust gas is collected by means of equipment based on United States Environmental Protection Agency (U.S. EPA) method 23A for stationary stack emissions. The concentrations of O2, CO, CO2, NO, NO2 and SO2 have also been measured. Six tests were carried out at 90-100 km/h on a route 100 km long. Two additional tests were done during the first 10 minutes and the following 60 minutes of the run to assess the effect of the engine temperature on PCDD/F emissions. The emission factors obtained for the vans varied from 1800 to 8400 pg I-TEQ/Nm3 for a 2004 model year van and 490-580 pg I-TEQ/Nm3 for a 2006 model year van. Regarding the passenger car, one run was done in the presence of a catalyst and another without, obtaining emission factors (330-880 pg I-TEQ/Nm3) comparable to those of the modern van. Two other tests were carried out on a power generator leading to emission factors ranging from 31 to 78 pg I-TEQ/Nm3. All the results are discussed and compared with literature.

BIDIRECTIONAL AC-DC CONVERTERS USING ONE CYCLE CONTROL (OCC) FOR ELECTRIC VEHICLES

The electric vehicle (EV) market has seen a rapid growth in the recent past. With an increase in the number of electric vehicles on road, there is an increase in the number of high capacity battery banks interfacing the grid. The battery bank of an EV, besides being the fuel tank, is also a huge energy storage unit. Presently, it is used only when the vehicle is being driven and remains idle for rest of the time, rendering it underutilized. Whereas on the other hand, there is a need of large energy storage units in the grid to filter out the fluctuations of supply and demand during a day. EVs can help bridge this gap. The EV battery bank can be used to store the excess energy from the grid to vehicle (G2V) or supply stored energy from the vehicle to grid (V2G ), when required. To let power flow happen, in both directions, a bidirectional AC-DC converter is required. This thesis concentrates on the bidirectional AC-DC converters which have a control on power flow in all four quadrants for the application of EV battery interfacing with the grid. This thesis presents a bidirectional interleaved full bridge converter topology. This helps in increasing the power processing and current handling capability of the converter which makes it suitable for the purpose of EVs. Further, the benefit of using the interleaved topology is that it increases the power density of the converter. This ensures optimization of space usage with the same power handling capacity. The proposed interleaved converter consists of two full bridges. The corresponding gate pulses of each switch, in one cell, are phase shifted by 180 degrees from those of the other cell. The proposed converter control is based on the one-cycle controller. To meet the challenge of new requirements of reactive power handling capabilities for grid connected converters, posed by the utilities, the controller is modified to make it suitable to process the reactive power. A fictitious current derived from the grid voltage is introduced in the controller, which controls the converter performance. The current references are generated using the second order generalized integrators (SOGI) and phase locked loop (PLL). A digital implementation of the proposed control ii scheme is developed and implemented using DSP hardware. The simulated and experimental results, based on the converter topology and control technique discussed here, are presented to show the performance of the proposed theory.