Welcome to join NextG2Com seminar on Interactive Programming of Performant Motions with Klas Nilsson, Amina Gojak and Philip Olhager from NextG2Com partner Cognibotics.

Physical motions by machines are increasingly being programmed, by machine designers, by system integrators, and/or by end-users. AI-driven robots is a well-known recent example, although it is being debated whether or not such robots will simply learn from demonstrations, but that is anyway yet another use-case for motion programming. However, there is up to now no such thing as motion programming within the discipline of computer science or IT in general. Cognibotics AB and Lund University bring a radical change to the current state of the art, including novel mechanics for motion physics not possible before, and the Julia-based (julialang.org) Juliet&Romeo motion programming technology that you now can learn more about on coursera.org/learn/juliet/. 

Real-time performance is combined with suitable abstractions for dealing with complexity and enabling code reuse. Code interpretation results in motions that change physical state, which is subject to human directives/adjustments/changes during runtime. Therefore, motion programming benefits greatly from interactive features such as backwards execution that also work physically on the factory floor (or domestically in the future), when for instance you are teaching a robot how to carry out the boring work.