The field of Robotics is developing rapidly around the world. Different types of robots such as humanoids, manipulators, and drones are in constant development by different formal research groups, university students and hobbyists.
Team-NUST was established formally in Pakistan, in 2013 with the aim of carrying out research in the rapidly progressing field of humanoid robotics, artificial intelligence, machine vision, motion planning, kinematics and navigation; and with the motivation to participate in RoboCup Standard Platform League and contribute towards achieving the 2050 goal of Humanoids defeating the Soccer World Cup Champions of that time.
The team has developed a comprehensive software architecture which takes the information and interactions from the environment and translates them to motion in an efficient and organized manner. The design is divided into layers, with selective synchronized information sharing between layers.
Kicking motion is a crucial component for soccer gameplay. Team-NUST has therefore developed an omni-directional kick for NAO with varying speeds. It is dependent on two main modules: trajectory generation and inverse kinematics. The team is currently working on two general areas for the goalkeeper to prevent the opponent team from scoring. The first area relates to creating and implementation of diving motions. The second area deals with positioning the goalkeeper in accordance with the situation on the field.
The team has also developed a machine vision system for the robot which spans areas including robot detection, goal post detection, field area extraction and corners (landmarks) detection on the image. All the vision algorithms were chosen for their robustness, resistance to lighting variations, and accuracy of results.
Accurately determining the position of robot in the field is necessary to allow better decision making. The determined landmarks are used by the sensor models for the Kalman filter and Particle filter, with both filters using an odometric model. The Particle filter is used to solve kidnap problems.
Intelligence exhibited by a robot comprises its intelligent awareness of the environment and predictive planning of its actions to support the team as a whole. The team applied a probabilistic processing algorithm on detected landmarks to provide a reliable stream of input to the system.
Team-NUST is a new team in RoboCup-SPL, and the only SPL-team from South Asia. The project is an immense inspiration and the team hopes to learn a lot from this experience, which will serve as a foundation for their future research work. With the exception of a few instances, the work by Team-NUST is original and will provide a fresh perspective to the currently faced challenges in RoboCup SPL.
More information about activities of Team-NUST is available at: http://rise.smme.nust.edu.pk/projects/RoboCupSPL/