People are always keen to discover the latest advancements in technology and to identify whether these can improve their daily lives. An initiative that made headlines was the 2007 DARPA Urban Challenge, which paved the way for further development in autonomous driving.

In the aviation industry, a number of intelligent systems, like the traffic collision avoidance system and the replacement of conventional manual flight controls with a fly-by-wire system, captivated technology enthusiasts. Financial forecasts are challenging as there is a great amount of uncertainty and noise in financial data. However, this challenge is being addressed through machine learning algorithms which can learn and predict using financial time series data.

Technology is also essential in different areas in the medical sector such as in the diagnosis of diseases and in patients’ intensive care. All these achievements depend on artificial intelligence. Through artificial intelligence, mobile robotic systems can actuate and interact with the surrounding environment. They can also be deployed to automate repetitive manual work or else to assist humans in difficult or unsafe scenarios.

I’ve always been captivated by the application of intelligent systems. After completing my BSc degree, I pursued my studies by reading for a Master’s in intelligent systems and robotics at the De Montfort University, Leicester, UK. My studies were spread over a three-year period on a part-time basis and focused on the core principles and techniques used in the development of an intelligent system. The subjects studied were intensive and varied from artificial intelligence programming based on Prolog, problem-solving, machine control through fuzzy logic, neural network concepts and practical implementations.

The module on evolution computing proved to be highly interesting as this helped me understand how genetic algorithms can be used to determine the optimal solution from a finite set of possible solutions. I also worked on a number of projects, such as the development of a mobile application, whereby artificial intelligence was employed to develop a shopping cart helping customers draw up a shopping list without much ado.

The Master’s degree also focused on different robotic aspects such as robot architecture, basic computer vision and robot control. Particular attention was given to robot mapping and localisation. I made treasure of the various techniques learnt throughout the different modules and employed them in my dissertation, which covered how artificial intelligence techniques can prove useful when used on robots in search and rescue operations.

My experiment sought to determine how homogeneous robotic systems can collaborate to carry out a search and rescue operation in a more efficient manner. I have always been fascinated by intelligent autonomous systems and how these can assist humans in dangerous situations.

I have always been fascinated by intelligent autonomous systems and how these can assist humans in dangerous situations

While the robots were built to showcase what I had learnt, my work was inspired by the role robots played when searching for victims in major man-made and natural disasters such as 9/11, the 2011 tsunami in Japan and the earthquake in Italy’s L’Aquila region.

I built the homogenous robots using Lego Mindstorms NXT. Although I found a number of limitations particularly with communication switching delays, the toolkit is one of the best solutions to start building robots for educational purposes, and it comes at an affordable price.

You can also introduce other mechanical parts such as gears which provide additional degrees of freedom. Although the physical design of the robot was important, my thesis focused primarily on the development of the software. I used the programming language leJOS throughout the development of the project. This Java-based language provides a much richer experience than the graphical programming environment NXT-G bundled with the NXT.

Since robots need to be autonomous, I implemented a set of techniques to ensure that the robot perceives its surrounding and takes the appropriate decision accordingly. For instance, fuzzy logic was used to control the steering and also its speed.

Each robot was able to operate autonomously within a predefined space simulating the inner structure of a house. The robots were designed to autonomously assist each other when trying to locate a trapped victim so as to cover vaster areas affected by a disaster within a short timeframe. They had to further collaborate between themselves to rescue victims and transport them out of the disaster area to a safe environment.

The robots were installed with sonar sensors to calculate the relative distances to objects representing obstacles as well as colour sensors to help them reach objects, which simulate the victims. They were also installed with motor encoders to estimate the position of the motors relative to the starting position.

The robots were engaged in a number of tasks such as avoiding obstacles and searching for and rescuing victims. Each had to prioritise its actions, such as choosing between tasks that necessitated immediate action and those that could follow suit, such as exploring unknown areas.

Since the sonars did not always return an accurate reading, a fuzzy controller was implemented to steer the robot out of danger. Depending on the distances recorded by each sonar, the controller suggested the best avoidance measure for the robot. Such measures included a change in the robot’s direction and steering. When a robot located and reached a victim, it lowered its robotic arm and opened the grip to lift the person, or requested the other robot to help it if it was not in a position to perform the task on its own.

The robots are not controlled by a human operator but are designed to autonomously navigate in an arena. The ‘brain’ of the robot ensures priority is given to urgent actions such as avoiding obstacles, while also communicating with the other robot through Bluetooth.

Since the robot’s memory space is very limited, certain complex tasks such as map building need to be carried out on a host computer. For this reason, both robots supply the host with sensory data on a regular basis.

In cases where one of the robots suffered a fault, the other was able to take over the pending tasks not to delay the search and rescue operation. This ensured that the goal was still achieved.

My studies were a challenging and interesting experience. At the end, I was awarded a distinction for my work.

I am now looking forward to further my studies in artificial intelligence and in other related areas.

The author’s MSc studies were partially funded by a Strategic Educational Pathways Scholarship (Steps) scholarship partly financed by the EU – European Social Fund under Operational Programme II – Cohesion policy 2007-2013, ‘Empowering People for More Jobs and a Better Quality of Life’.