Module Check: How Robots Decide
You have traveled through the full landscape of robot intelligence and autonomy — from the basic heartbeat of the perceive-plan-act cycle to the sophisticated challenge of making decisions under uncertainty, and from the spectrum of human control to the frontier of robots that learn from experience. Before you move on, this module check gives you a chance to consolidate everything, make sure the key ideas stick, and apply them together in one final challenge.
Flashcards — click each card to reveal the answer
Module Quiz
A robot navigating a grocery store detects a spilled liquid it was not expecting and recalculates its path around it while continuing its delivery task. Which stages of the perceive-plan-act cycle are illustrated in this sequence?
A robot starts with a belief that it is in one of twelve possible locations in a building. After scanning a distinctive landmark with its camera, its probability distribution collapses to just two possible locations. After a second scan reveals a unique feature, it is nearly certain of its location. What technique is this robot using?
A company is deploying a robot to autonomously manage a nuclear waste storage facility. The environment is hazardous for humans to enter, the tasks are repetitive and well-defined, and communication latency with the surface makes teleoperation unreliable. What level of autonomy is most appropriate?
A robotics team trains a walking robot entirely in simulation. In the simulation the robot walks perfectly. On the real robot it stumbles on every floor surface. The team applies domain randomization — randomly varying floor friction, terrain roughness, and motor delays during simulation training. After retraining, the real robot walks reliably. What problem did domain randomization solve?
During a warehousing mission, a robot's right-side distance sensor suddenly reports an impossible reading of zero meters in open space. The robot slows down, switches to relying on its left-side sensor and camera, and continues the task at reduced speed. It also flags a maintenance alert. Which robustness concepts are demonstrated?
Synthesis Challenge
The Complete Autonomous System
- FINAL CHALLENGE: You will design a complete autonomous robot system that integrates every major concept from this module.
- THE SCENARIO: A city is deploying an autonomous robot to inspect park infrastructure — checking benches, trash cans, water fountains, and lighting fixtures for damage or graffiti — and sending photo reports to the parks department. The robot will operate in a public park from 6 AM to 8 AM before the park opens, but joggers and dog walkers may still be present.
- YOUR DESIGN MUST ADDRESS ALL SIX AREAS:
- 1. PERCEIVE-PLAN-ACT: Describe the three stages of the robot's cycle for a single inspection stop (approaching one bench, inspecting it, and moving on).
- 2. PATH PLANNING: What mapping approach will you use? How will the robot handle a path blocked by a parked bicycle it did not know about?
- 3. UNCERTAINTY: Name two sources of uncertainty the robot will face during a dawn inspection (hint: think about lighting conditions). How should it handle them?
- 4. AUTONOMY LEVEL: What level of autonomy is appropriate? The parks department wants daily automated reports but cannot staff a human operator at 6 AM. Justify your answer.
- 5. ROBUSTNESS: Describe two faults that could occur and the recovery behavior for each. Also describe how the system degrades gracefully if the main camera fails at the start of the inspection run.
- 6. SIMULATION: Before deployment, what scenarios would you include in a simulation test suite? Name at least four. What real-world tests would you run after simulation before going live in the park?
- Write up your complete design. Be specific — avoid vague answers like 'the robot will handle it.' Name the algorithms, behaviors, and design choices and explain why you chose each one.