3. Control

Question 1

What are the principles of deliberative and reactive control? Name some examples where they are used

Answer 1

Deliberative - slow but strong in strategical searching and planning

- Principle = think hard, act later!
- Dates back to early days of AI (first "thinking" robot Shakey in 1960s)
- Used in chess software (aka any software where route-planning occurs)

Reactive - fast reaction instead of deep thinking

- Principle = **don't think, react!** 
- Uses a set of condition-action rules that couple action to specific sensed situations

Question 2

What are the modules of a navigation system and their roles?

Question 3

What is open-loop control? (principle, block diagram) and name some examples

Answer 3

Open-loop - no connection between actuation effect and user’s goal

- Controller only looks at its input signal in order to decide what to do with the actuator. Once an input demand has been made, the system is left to respond in the hope that it will respond correctly
- Ex: traffic light recognition in an autonomous car

    - If controller recognizes a red traffic light, it will give the stop command to the wheels' actuator without actually checking whether the breaks really worked
- Other exs: microwave ovens and programmable light switch timers

Block Diagram:
input control signal -> controller -> actuator -> output controlled variable

Question 4

What is feedback closed-loop control? (principle, block diagram) and name some examples.

Answer 4

Closed-loop - user inputs desired goal directly into system

• How does it work?
  
  • Control system continuously compares current state to desired state and calculates the error by subtracting the output from the goal
    
    • Depending on sign of error, current state is adjusted to minimize error
• Feedback control system
  
  • Output controlled variable is sensed, fed back to the input, and compared with desired goal
• Examples
  
  • Cruise, climate and transmission control systems in smart cars, fly-by-wire flight control systems, robotic surgery, network congestion and chemical processes control systems

Block Diagram:

Goal (desired value) -> Comparator -> Controller -> Actuator -> Output (controlled variable) OR Sensor (error) -> Comparator (repeat)

Question 5

How can a smart vehicle detect obstacles? (sensors principles). Write some simple controllers in pseudocode

Answer 5

Different handling of obstacles

• Detected and avoided - in autonomous vehicles
• Detected and retrieved - in earthquake areas or fire sites
• Detected and tracked - tracking intruders, guided museum tours

Pseudocode:
let vehicle drive until it bumps into obstacle: 
while (no contact) 
drive-straight-forward 
end_while 
stop driving
stop vehicle 
before it bumps into obstacle: while distance > 20 cm 
drive forward 
end_while 
stop the motors

Question 6

How to follow a black line? sensors, controllers in pseudocode

Answer 6

windshield camera or laser scanners are used

Pseudocode:
while (true)
if (both sensors detect black) then stop driving break
while
else if (both sensors detect white) then
drive straight forward
else if (left sensor sees black) then turn left
else if (right sensor sees black) then turn right
end while