Alvin Y. Chua

Vision based pedestrian detection using Histogram of Oriented Gradients, Adaboost & Linear Support Vector Machines

Pedestrian detection systems are valuable in a variety of applications such as in advanced driver assistance systems and advanced robots. This study presents a pedestrian detection system that uses Histogram of Oriented Gradients (HOG) as feature descriptor, and AdaBoost and Linear Support Vector Machines (SVM) as classifiers. The entire system is tested and evaluated in both publicly available databases and personally acquired videos. The pedestrian detection system has been tested and results show that it can detect pedestrians. Experiments showed that the system is up 20% faster compared to OpenCVs default detector.

Identification of river hydromorphological features using Viola-Jones Algorithm

In this paper, a quadcopter equipped with a camera was used to capture images from a river. These captured images were used as training data in the detection program used to detect the hydromorphological features in the area of the river such as trees, roofs, roads and the shore. The Viola-Jones Algorithm was used in order to detect, identify and recognize hydromorphological features due to its speed and simplicity of implementation. Testing was done using different images to verify the effectiveness of detection. System evaluation and success of the appropriateness of the Viola-Jones Algorithm was determined using the percentage of correct detected features in the image. The study showed that the Viola-Jones has shown that it is effective in detecting some features due to the complexity of the hydromorphological images.

Micro and Nano Silver-Graphene Composite Manufacturing via Horizontal Vapor Phase Growth (HVPG) Technique

In this study, nanosilver-graphene composites were successfully manufactured via the horizontal vapor phase growth (HVPG) technique. A quartz tube loaded with the starting material, equal masses silver (Ag) powder and multi-layer graphene (Ge), was evacuated to ~10-6 Torr, sealed, and then baked at 1200°C for 6 hours, with its orientation such that a horizontal temperature gradient was generated across the tube. Scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) analysis revealed variations in the structure and composition of the nanomaterials deposited on different regions of the tube, and the diameter of the nanomaterials was found to decrease with decreasing temperature.

Statistical Design Analysis of Silver-Titanium Dioxide Nanocomposite Materials Synthesized via Horizontal Vapor Phase Growth (HVPG)

This study aims to create a statistical design analysis of the synthesized Silver-Titanium dioxide nanocomposite materials using HVPG technique. The analysis was obtained to report the nature of nanomaterials produced for a specific growth time and temperature. The synthesized nanocomposite materials were characterized using the Scanning Electron Microscope and Energy Dispersive X-ray system. Using the JMP ANOVA Software, the experimental results revealed an empirical equation to predict the behavior of the synthesized material formed at different growth zones.

Quadrotor Research Platform with Obstacle Avoidance

This study focuses on the development of a quadrotor research platform with effective obstacle avoidance system using Bug Algorithm. The quadrotor system consist of a DJI frame with a Crius All-in-One Pro as a flight controller with an Arduino Mega2560 to process the data from the sensor. It is also equipped with four ultrasonic range finders to provide the sensing for collision avoidance. Bug Algorithm was chosen due to its simplicity and effectiveness in implementation of obstacle avoidance applications. The results shows that the quadrotor system is able to perform basic flight maneuvers and that the collision avoidance system was able to protect the quadrotor from hitting objects in its environment.

Random execution of a set of contacts to solve the grasping and contact uncertainties in robotic tasks

This paper addresses the problem of identifying the uncertainties present in a robotic contact situation. These uncertainties are errors and misalignments of an object with respect to its ideal position. The paper describes how to solve for the errors caused during grasping and errors present when coming into contact with the environment. A force sensor is used together with Kalman Filters to solve for all the uncertainties. The straightforward use of a force sensor and the Kalman Filters is found to be effective in finding only some of the uncertainties in robotic contact. The other uncertainties form dependencies that cannot be estimated in this manner. This dependency brings about the problem of observability. To make the unobservable uncertainties observable a sequence of contacts are used. The error covariance matrix of the Kalman Filter (KF) is used to obtain new contacts that are required to solve for all the uncertainties completely. There is complete freedom in choosing which unobservable quantity to be excited in forming the next contact. The paper describes how these new contacts can be randomly executed. A two dimensional contact situation will be used to demonstrate the effectiveness of the method. Experimental data are also presented to prove the validity of the procedure. Due to the non-linear relationship between the uncertainties and the forces, an Extended Kalman Filter (EKF) has been used.

Random contact strategy using the Kalman filter to solve the robotic contact uncertainty problem

This paper addresses the problem of finding the uncertainties present in a robotic contact. There are two kinds of uncertainties: grasping uncertainties and contact uncertainties. The grasping uncertainty vector contains errors (angles and displacements) associated with improper grasping. The contact uncertainty vector contains errors in angles and positions of nominal contact. A force sensor is used together with Kalman filters to solve the uncertainties. The straightforward use of Kalman filters is found to be effective in finding only some of the uncertainties. The quantities that form dependencies cannot be estimated in this manner. This dependency brings about the problem of observability. The unobservable quantities can be determined using a sequence of contacts. The error covariance matrix of the Kalman filter can indicate the directions of dependency and accuracy of the values estimated. A new contact in any of the dependent directions can be randomly chosen as the next contact to try. The relational transformations between contacts are used to eventually obtain the complete solution. A two dimensional contact situation will be used to demonstrate the effectiveness of the method. Experimental data are also presented to prove the validity of the procedure. Due to the non-linear relationship between the uncertainties and the forces, an extended Kalman filter (EKF) has been used.