MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance

As an effective means of solving collision problems caused by the limited perspective on board, the cooperative roadside system is gaining popularity. To improve the vehicle detection abilities in such online safety systems, in this paper, we propose a novel multi-sensor multi-level enhanced convolutional network model, called multi-sensor multi-level enhanced convolutional network architecture (MME-YOLO), with consideration of hybrid realistic scene of scales, illumination, and occlusion. MME-YOLO consists of two tightly coupled structures, i.e., the enhanced inference head and the LiDAR-Image composite module. More specifically, the enhanced inference head preliminarily equips the network with stronger inference abilities for redundant visual cues by attention-guided feature selection blocks and anchor-based/anchor-free ensemble head. Furthermore, the LiDAR-Image composite module cascades the multi-level feature maps from the LiDAR subnet to the image subnet, which strengthens the generalization of the detector in complex scenarios. Compared with YOLOv3, the enhanced inference head achieves a 5.83% and 4.88% mAP improvement on visual dataset LVSH and UA-DETRAC, respectively. Integrated with the composite module, the overall architecture gains 91.63% mAP in the collected Road-side Dataset. Experiments show that even under the abnormal lightings and the inconsistent scales at evening rush hours, the proposed MME-YOLO maintains reliable recognition accuracy and robust detection performance.

A Noncontact Method for Calibrating the Angle and Position of the Composite Module Array

Freeform surface is one of the research focuses in the measurement field. A composite module is composed of a plane and rotating paraboloid. The composite module array can identify 21 geometric errors of the machine tool in a wide range, which is composed of several composite modules. Eliminating the error of the array itself is of great significance for improving measurement accuracy. For this reason, this paper proposed a noncontact method for calibrating the angle and position of the composite module array. This paper used a self-developed angle sensor to access corresponding information and established the mathematical model according to the freeform surface’s geometric characteristics to achieve calibration. In addition, the influence of array placement error on calibration was analyzed. The experimental results showed that the angle repeatability was within 0.4″ around the X-axis and within 0.3″ around the Y-axis, and the position repeatability was within 0.4 µm in the X direction and within 0.7 µm in the Y direction. The measurement comparison experiments with high-precision laser interferometer and uncalibrated array verified the correctness of the experimental results. This method provides an important reference for practical application and freeform surface array calibration, and creates conditions for the implementation of machine tool error detection.

An optimization scheme for a multilayer armour module against 7.62 mm armour piercing projectile

This study presents a methodology to find the optimal sequence and thicknesses of individual material layers in a multilayer armour module. The methodology is demonstrated with application to three different metal alloys: Armox-500T, Ti-6Al-4V and Al-2024. Numerical simulations are performed first to study the ballistic impact behaviour of these three materials using AUTODYN-3D code. The results of numerical simulations are compared with experimental results for validating the numerical models. Thereafter, a three-layer armour module consisting of these three materials is optimized to defeat 7.62 armour piercing projectile with minimum weight. The optimization process involves carrying a set of numerical simulations based on the design of experiment approach to generate a response surface for the ballistic performance of a composite module. A new ballistic performance parameter is introduced to measure the ballistic response of the module by combining depth of penetration and residual velocity of the projectile to bring uniformity between two cases of partial and complete penetration. The proposed parameter provides more information on ballistic performance. The response surface for ballistic performance parameter is generated in terms of thicknesses for six possible combinations of three material layers. The adequacy of the proposed optimization scheme is confirmed with ballistic experiments. The sequence Armox-500T/Ti-6Al-4V/Al-2024 with thicknesses 5.5, 8.5 and 13 mm, respectively, is found to be the best against 7.62 mm armour piercing projectile. Furthermore, the performance of each individual material is compared with an optimized three-layer armour module. The composite module is found to be weight efficient over Armox-500T, Al-2024 and provides better thickness efficiency over Al-2024. The weight efficiency and thickness efficiency of Ti-6Al-4V are found to be comparable to the composite module. This study emphasizes the necessity of developing new procedures to provide reliable estimates of design parameters for a multilayer armour module.

Implementation of an OBD-II diagnostics tool over CAN-BUS with Arduino

Since its beginnings, the goal of On Board Diagnostics [OBD] standard has been to take control over gasses emission in automotives and its effects in the environment. This project implements a system based on OBD-II diagnostic protocol over a CAN bus, which allows both, displaying real time variables and diagnostic the vehicle status to get codes about its functioning, failures and energy efficiency. On board diagnostic systems allow to obtain failure codes stored and several relevant variables as: speed, gas level and CO2 emission levels, in real time. An OBD-II system centered at the end of the bus, corresponding to the scanner or diagnostic unit, was implemented on an Arduino Mega 2560 board connected to a CAN transceiver-controller composite module. The scanner has a USB connection that makes it easy to view the data recovered in a versatile way on a PC through a graphical interface created in LabVIEW ™.

Modular Design Concept of a Humanoid Robot Platform

Most of the biped robots or humanoid robots developed so far have been designed with serialactuated mechanism. the ratio between power and weight has not been very good. Some abilities of the humanoid robot are dependent on the weakest motor just like the cannikin law. Most of the human joints are not serial-actuated but parallel-actuated except the knee joint which is a one DOF joint. In the study, two kinds of module are put forward: basic module and composite module, and the configurations of the joints and limbs are designed by these modules. Based on logic diagram, many kinds of the humanoid robot are structured with these joints and limbs. As an applicable example, the humanoid robot SJTU-HR1 is presented.