Reactive Behavior vs Planned Behavior

A robot can respond directly to what it senses, or it can reason about a sequence of future actions.

Both approaches are useful.

Reactive behavior

A reactive rule might be:

if obstacle_distance < safe_distance:
    stop

The response is local and immediate. It does not require a map of the entire mission.

Reactive behavior is useful for:

• safety boundaries;
• obstacle avoidance;
• simple tracking;
• local corrections.

Planned behavior

A planner considers how to reach a goal through a sequence of states or positions.

For a grid map:

S . . #
# # . #
. . . G

The robot may search for a route from S to G.

Planning becomes important when one locally good action can create a bad future state.

Why real robots mix them

A planned path cannot predict every disturbance.

A practical architecture might use:

mission goal
   ↓
global path plan
   ↓
local motion target
   ↓
reactive obstacle / safety layer
   ↓
actuator control

The planner chooses where to go. Reactive layers protect or adjust the motion using current evidence.

Failure mode: fighting controllers

If a planner commands forward motion while a safety layer commands stop, which one wins?

Priority must be explicit.

A system where two subsystems can command the same actuator without an ownership rule becomes unpredictable.

Compare architectures

For a robot navigating a hallway, write one behavior that should be:

• reactive;
• planned;
• human-controlled.

Explain why.

Then draw the command-priority order when those behaviors disagree.

Reactive behavior can be extremely good

A line follower that steers from current sensor error is reactive. It does not need a map of the entire course to make the next steering correction.

Reactive behavior is strong when:

• the needed evidence is local and current;
• the environment changes quickly;
• a short response time matters;
• long-horizon optimization is unnecessary.

Planning becomes useful when current action depends on future consequences.

A warehouse robot may need to choose among several routes, reserve a narrow aisle, or avoid a path that will become blocked.

Hybrid systems use both

Real robots often combine them.

global planner: choose corridor sequence
        ↓
local behavior: follow path and avoid obstacle
        ↓
low-level control: track velocity

If an unexpected cart appears, waiting for a full global re-plan before slowing down would be foolish. Local reactive logic can protect the immediate motion while a higher layer decides what to do next.

Compare failure modes

Do not ask which approach is "more intelligent." Ask what information the task requires and how quickly the system must react.