Trustworthiness and Safety for Intelligent Ethical Logical Agents via Interval Temporal Logic and Runtime Self-Checking