Detecting pedestrians in autonomous driving is a safety-critical task, and the decision to avoid a a person has to be made with minimal latency. Multispectral approaches that combine RGB and thermal images are researched extensively, as they make it possible to gain robustness under varying illumination and weather conditions. State-of-the-art solutions employing deep neural networks offer high accuracy of pedestrian detection. However, the literature is short of works that evaluate multispectral pedestrian detection with respect to its feasibility in obstacle avoidance scenarios, taking into account the motion of the vehicle. Therefore, we investigated the real-time neural network detector architecture You Only Look Once, the latest version (YOLOv4), and demonstrate that this detector can be adapted to multispectral pedestrian detection. It can achieve accuracy on par with the state-of-the-art while being highly computationally efficient, thereby supporting low-latency decision making. The results achieved on the KAIST dataset were evaluated from the perspective of automotive applications, where low latency and a low number of false negatives are critical parameters. The middle fusion approach to YOLOv4 in its Tiny variant achieved the best accuracy to computational efficiency trade-off among the evaluated architectures.

Hydrophobization of cotton fabrics was carried out with the use of bifunctional polysiloxanes with various contents of functional groups. Polysiloxanes contained in their structure groups capable of bonding to substrates (trialkoxysilyl or glycidyl ones) and fluoroalkyl groups showing surface activity. Two methods of surface modification were compared: (1) a one-step method via the chemical modification of fabrics with solutions of bifunctional polysiloxanes and (2) a two-step method—via preliminary modification of fabrics with silica sol followed by chemical modification with solutions of bifunctional polysiloxanes. The hydrophobicity was determined by measuring the water contact angle by drop profile tensiometry. Changes in the surface topography were examined by scanning electron microscopy. Superhydrophobic fabrics were obtained by a simple one-step method by the chemical modification in solutions of bifunctional polysiloxanes. The fabrics maintained their superhydrophobic properties even after multiple washings. The modification does not cause any changes visible to the naked eye, such as stiffening, color changes or a decrease in mechanical properties.

Adaptive multi-legged walking robots are predestined to be applied in rough and hazardous terrain. Their walking and climbing skills allow them to operate at places that are unreachable for most wheeled vehicles. In this paper, we present the design and development of the new six-legged walking robot LAURON V with its improved kinematics and robust mechanical structure. Each leg has four independent joints that enable LAURON to cope with steep inclines, large obstacles and makes it possible to manipulate objects with its front legs. Autonomy, robustness and a large payload capacity together with its impressive terrain adaptability make LAURONV highly suitable for all kinds of field applications.

The blends based on poly(lactic acid), (PLA) and polyacrylates (1:1 weight ratio) have been prepared by photopolymerization of three different multifunctional acrylate monomers (pentaerythritol triacrylate, pentaerythritol tetraacrylate and dipentaerythritol pentaacrylate) in matrix of linear PLA. The structure of these systems has been studied by Atomic Force Microscopy (AFM), Differential Scanning Calorimetry (DSC) and X-ray Diffraction (XRD). It was found that the blends are highly crosslinked, rough, heterogeneous and some of them – semicrystalline. The effect of UV-irradiation on surface morphology of obtained mixtures was studied. The viscometry has been applied for measurements of changes of PLA molecular weight resulting of photodegradation. It was shown that blends are much more photostable than pure PLA.