Artificial intelligence for intelligent robots - SPECTRO

Artificial Intelligence for Intelligent Robots
This course explores how AI enables robots to perceive, learn, move, and interact intelligently with the world. It covers key AI technologies used in robotics, including computer vision, machine learning, reinforcement learning, motion planning, and human-robot interaction, with a strong focus on real-world applications and hands-on experimentation.

Module descriptions

Module 1: What makes a robot intelligent?

This module introduces AI-powered robotics, distinguishing between classical and AI-driven robots. It delves into levels of autonomy, from teleoperated to fully autonomous systems, and addresses ethical considerations in deploying intelligent robots. Real-world examples include self-driving cars, humanoid robots, and warehouse automation systems.​

Hands-on: compare rule-based vs. AI-driven robot behavior in a virtual simulation.​

Module 2: How robots perceive the world – computer vision & sensors

Intelligent robots utilize various sensors, such as cameras and LiDAR, to interpret their surroundings. This module explains sensor fusion, deep learning-based perception, traditional computer vision techniques, sensor calibration, and data preprocessing. Challenges of perception in dynamic environments are also explored.​

Hands-on: implement an object detection AI model to understand how robots identify objects.​

Module 3: Robot learning – how machines adapt and improve

This module covers how robots learn through machine learning and reinforcement learning, enabling them to recognize patterns, improve through trial and error, and adapt to environments without human intervention. Topics include transfer learning and the role of simulation environments in training robotic systems.​

Hands-on: train a reinforcement learning-based robot in a simulation environment.​

Module 4: Motion planning and navigation – AI on the move

Exploring how AI assists robots in navigating and avoiding obstacles, this module covers path planning algorithms (A*, Dijkstra), Simultaneous Localization and Mapping (SLAM), probabilistic roadmaps, rapidly-exploring random trees (RRTs), and real-time navigation challenges in unstructured environments.​

Hands-on: simulate a robot navigating a dynamic environment using various planning algorithms.

Module 5: Human-robot interaction – can AI make robots social?

AI enables robots to understand speech, recognize emotions, and interact naturally with humans. This module explores speech recognition, natural language processing (NLP), social robots in customer service and healthcare, design of intuitive user interfaces, and case studies of social robots in different cultural contexts.​

Hands-on: test an AI chatbot and voice recognition system to evaluate human-robot interaction.​

Module 6: AI for swarm robotics – when robots work together

Swarm intelligence allows multiple robots to collaborate effectively. This module explains how AI enables self-organizing drone swarms, warehouse automation, rescue robotics, bio-inspired algorithms driving swarm behaviour, and scalability challenges in deploying large robotic swarms.​

Hands-on: experiment with multi-agent robotic coordination in a simulation environment.​

Module 7: Ethical & real-world challenges in AI robotics

AI-powered robots present safety, fairness, and ethical concerns. This module explores AI transparency, risks of autonomous robots, the impact of AI robotics on employment and the economy, and the importance of transparency and explainability in AI-driven robots.​

Hands-on: participate in a debate on AI ethics in robotics, focusing on real-world scenarios.​

Final course takeaways

• AI enables robots to perceive, learn, move, and interact autonomously.​
• Various AI techniques (ML, DL, RL) power diverse robotic applications.​
• Computer vision, NLP, and reinforcement learning are pivotal in shaping the future of robotics.​
• AI-powered robots offer exciting possibilities alongside ethical considerations.

About SPECTRO

This course is delivered by SPECTRO. SPECTRO is a consortium consisting of 12 higher education institutions from 7 different countries, 2 innovative SMEs, and one leading research center in Information Systems and EIT Digital. SPECTRO is co-funded by the European Union’s Digital Europe Programme.