Robot navigation

Goal: to enable robots in manufacturing to perform autonomous mapping, localisation and navigation
Use case

JinZone

Implemented algorithms

Environment Reconstruction

Algorithm for reconstructing objects based on images - SIFT (Scale-invariant feature transform)

Optimum route calculation

Algorithm A* for calculating and programming optimal routes

Semantic reasoner

Elif Lab's proprietary reasoner and controller, capable of harmoniously handling information from movements and external stimuli and dynamically and adaptively generating sequences of actions to achieve the desired result

Explainability

Semantic component that enables transparent communication of robot choices and allows the robot to incorporate operator and user inputs

Interested in learning more about our approach to artificial intelligence and software development for robotics?
Visit our dedicated website: https://robotics.eliflab.com

JinZone

JinZone is a project that aims to create a service capable of supporting robots within the manufacturing sector, in the mapping, localisation and navigation phases.

It combines established artificial intelligence techniques with a proprietary reasoner and orchestrator layer developed by Elif Lab, which exploits established models of the mind in psychology and philosophy and their proprietary mathematical-algorithmic formalisation to create a new AI paradigm.

In this way, Jinzone retains the advantages of known navigation technologies, but can exploit an additional layer of abstraction, reasoning and control that operates – like the model of the mind – on abstract representations.

This adaptive technology allows for better orientation of activities without costly re-training, to operate in dynamic and chaotic environments and to have more control over all phases.

Decisions are made by the robot are transparent and can be evaluated and redirected. In addition, JinZone makes it easier than other technologies to operate in unfamiliar or unmapped environments, reducing the time it takes to introduce robots into new contexts.

 

  • Client server
  • MQTT protocol

JinZone has been selected in the VOJEXT open call and is currently funded in the context of the European Horizon 2020 research and innovation programme, grant agreement No 952197.