Body of Knowledge

The Body of Knowledge helps to support the development of safe autonomous systems by providing practical guidance on assurance and regulation.

Use the coloured boxes below to explore the Body of Knowledge. For each assurance / regulatory consideration you will find:

  • a definition of the assurance objectives (what must be demonstrated when putting a RAS into operation)
  • a description that provides further information and rationale for the assurance objectives
  • practical guidance on how to meet the assurance objectives (you may find multiple approaches for demonstration to reflect different technologies or domains)

Every assurance consideration already has a definition and description. The guidance is being added as it is developed and written. You can use the links below to go straight to some of the guidance that is already in place:

1.1.2 - practical guidance for 'defining the operating environment' (automotive domain)

1.1.3 - practical guidance for 'defining operating scenarios' (automotive domain)

1.2 - practical guidance for 'identifying hazardous system behaviour' (automotive domain)

1.2.1 - practical guidance for 'considering human/machine interactions' (social care domain)

2.3 - practical guidance for 'implementing requirements using ML' (cross-domain)

2.3.1 - practical guidance for 'sufficiency of training' (cross-domain)

2.3.2 - practical guidance for 'sufficiency of the learning process' (cross-domain)

2.3.3 - practical guidance for 'verification of the learned model' (cross-domain)

2.8 - practical guidance on interpretability techniques for 'explainability' (cross-domain)

2.8 - practical guidance on interpretability requirements for 'explainability' (cross-domain)

2.8 - practical guidance on interpretability evaluation for 'explainability' (cross-domain)

Section 1

Section 2

Section 3

Section 4

 

Downloads: