The group will be represented by Prof. Bruce MacDonald and Dr Pau Medrano-Gràcia from The University of Auckland and Prof. Mike Duke from the University of Waikato. The seminar will summarize their research partnership to create robotics technology for kiwifruit and apple orchards, with the Universities of Auckland and Waikato, New Zealand’s Plant and Food Research Institute, and New Zealand company Robotics Plus. This includes an autonomous mobile robot, a targeted pollination system for kiwifruit flowers, a four-arm harvester for kiwifruit, and a prototype apple harvester.
The performance of autonomous robots, i.e. robots that can make their own decisions and choose their own actions, is becoming increasingly impressive, but most of them are still constrained to labs, or controlled environments. In addition to this, these robots are typically only able to do intelligent things for a short period of time, before either crashing (physically or digitally) or running out of things to do. In order to go beyond these limitations, and to deliver the kind of autonomous service robots required by society, we must conquer the challenge of combining artificial intelligence and robotics to develop systems capable of long-term autonomy in everyday environments. This talk will present recent progress in this direction, focussing on the mobile robots for security and care domains developed by the EU-funded STRANDS project (http://strands-project.eu) which have so far completed over 106 days of autonomy in real service environments. In particular the presentation will cover our approach which combines probabilistic verification and machine learning to produce a planning system which controls how the robots select and execute their tasks over these extended periods of autonomy.
Nick Hawes is an Associate Professor of Engineering Science in the Oxford Robotics Institute at the University of Oxford. His research applies techniques from artificial intelligence to allow robots to perform useful tasks for, or with, humans in everyday environments (from moving goods in warehouses to supporting nursing staff in a care home). He is particularly interested in how robots can understand the world around them and how it changes over time (e.g. where objects usually appear, how people move through buildings etc.), and how robots can exploit this knowledge to perform tasks more efficiently and intelligently.
While reinforcement learning has led to promising results in robotics, defining an informative reward function often remains challenging. In this talk, I will give an overview about different reward learning approaches and how they can be used for learning robotics policies in practice. In particular, I will present an efficient hierarchical reinforcement learning approach for learning how to grasp objects from preferences. Furthermore, I will show how inverse reinforcement learning can be used to learn flocking behavior of birds, which could potentially be used for apprenticeship learning of robot swarms.