Self-Assessment Quiz: Guidance, Navigation, and Control

Twenty questions to check your grasp of the GN&C loop, state estimation, guidance, the PID controller, and autonomy. Answer each before opening the key. Aim for 16 or more.

Question 1

In the GN&C loop, the box that answers "where am I, and how sure am I?" is:

A) guidance B) navigation C) control D) actuation

Question 2

The essential difference between closed-loop and open-loop control is that closed-loop control:

A) uses a bigger computer B) measures the result and feeds it back to correct the error C) requires no sensors D) runs only on the ground

Question 3

A rocket cannot be flown reliably open-loop (no feedback) mainly because:

A) it would use too much fuel B) unmodeled disturbances accumulate with nothing to cancel them C) the rocket equation forbids it D) radio signals are too slow

Question 4

Navigation is best described as:

A) choosing the target orbit B) real-time state estimation from noisy sensors plus a dynamical model C) firing the thrusters D) computing the pitch program

Question 5

In the Kalman filter used for navigation, the estimate's uncertainty:

A) grows in the update step, shrinks in the predict step B) grows in the predict step, shrinks in the update step C) never changes D) always grows

Question 6

An IMU (inertial measurement unit) must be paired with an absolute sensor because:

A) it is too heavy to use alone B) integrating its acceleration and rate outputs lets bias accumulate into unbounded drift C) it only works in sunlight D) it measures position directly and needs no help

Question 7

Powered explicit guidance (PEG) for ascent is valuable because it:

A) follows a fixed stored trajectory exactly B) recomputes the optimal steering and cutoff each cycle to hit the target despite dispersions C) eliminates the need for a control loop D) works only below the atmosphere

Question 8

The "velocity-to-be-gained" is:

A) the current speed of the vehicle B) the vector difference between the velocity needed at the target and the current velocity C) the exhaust velocity of the engine D) the escape velocity

Question 9

In the PID law $u = K_p e + K_i\int e\,dt + K_d\,\dot e$, the term that eliminates steady-state error against a constant disturbance is the:

A) proportional term B) integral term C) derivative term D) none — steady-state error is unavoidable

Question 10

The derivative term in a PID controller mainly:

A) removes steady-state droop B) adds damping and reduces overshoot, at the cost of noise sensitivity C) increases the steady-state gain D) integrates past error

Question 11

A pure proportional controller holding against a constant disturbance settles at a nonzero "droop" because:

A) the sensor is biased B) it produces a correction only when an error is present, so it holds just enough error to generate the opposing effort C) the integral term saturates D) the actuator is broken

Question 12

For the discrete loop $\theta_{n+1} = (1-K_p)\theta_n + 1$, the loop is stable only for:

A) $K_p > 2$ B) $0 < K_p < 2$ C) any $K_p$ D) $K_p < 0$

Question 13

Thrust vector control (TVC) steers a launch vehicle by:

A) firing reaction wheels B) gimbaling the engine so its thrust line points slightly off the center of mass, creating a torque C) using magnetorquers against Earth's field D) shifting the payload

Question 14

GN&C loops are usually nested (cascaded) so that:

A) the vehicle uses less propellant B) a fast inner loop (rates/attitude) runs inside a slower outer loop (guidance), each simple enough to be stable C) the ground can take over D) the star tracker updates faster

Question 15 (True/False, justify)

"Making the proportional gain $K_p$ very large is the best way to remove a control loop's steady-state droop." True or false? Justify in one sentence.

Question 16 (True/False, justify)

"Because the fused Kalman uncertainty uses inverse-variance addition, it can be smaller than both the prediction's and the measurement's uncertainty." True or false? Explain.

Question 17 (True/False, justify)

"A Mars lander could be flown down by mission controllers on Earth watching live telemetry." True or false? Say why, with the relevant timescales.

Question 18 (Short answer)

Name the four stages of one trip around the GN&C loop (four verbs), and give one piece of real hardware associated with the first and the last.

Question 19 (Short answer)

A navigation filter predicts a position with $\sigma_{\text{pred}}^2 = 16$ and gets a measurement with $\sigma_{\text{meas}}^2 = 4$. Compute the Kalman gain and the updated variance.

Question 20 (Short answer)

In one or two sentences, state the inequality (in words) that forces a mission to be autonomous, and give the canonical example from this chapter.


Answer Key

Q Ans Note
1 B Navigation estimates the state and its uncertainty.
2 B Feedback: measure the output, compute the error, correct.
3 B Disturbances (wind, slosh, thrust offset) accumulate unchecked without feedback.
4 B Blend a dynamical model with noisy measurements, in real time, onboard.
5 B Predict grows uncertainty (process noise); update shrinks it (information).
6 B Integrating acceleration/rate turns bias into unbounded drift; needs absolute fixes.
7 B Closed-loop: recompute optimal steering and cutoff to hit the target despite dispersions.
8 B The vector from current velocity to the velocity needed at the target.
9 B The integral accumulates error until it is exactly zero.
10 B Derivative responds to error rate — damping/anticipation — but amplifies noise.
11 B P outputs only when an error exists, so it holds a standing error to fight a constant disturbance.
12 B $|1-K_p|<1 \Rightarrow 0
13 B Gimbaling offsets the thrust line from the c.g., producing a steering torque.
14 B Fast inner loop inside a slow outer loop; each provably stable and modular.
15 False High $K_p$ destabilizes the loop (past $K_p=2$ here); the correct cure for droop is the integral term.
16 True $1/\sigma_+^2 = 1/\sigma_{\text{pred}}^2 + 1/\sigma_{\text{meas}}^2$, so the fused variance is below either input.
17 False Mars round-trip light-time (~28 min) far exceeds the ~7-min descent; the ground sees a 14-min-old recording. It must be autonomous.
18 Sense, estimate, decide, actuate. First: a sensor (IMU/star tracker/radar). Last: an actuator (TVC/thruster/reaction wheel).
19 $K = 16/(16+4) = 0.8$; updated variance $= (1-0.8)(16) = 3.2$ (so $\sigma^+ = 1.79$).
20 When round-trip light-time exceeds the reaction time available, control must be onboard; canonical example: the Mars EDL "seven minutes of terror."

Topics to review by question

Questions Topic Section
1, 18 The loop and its boxes §27.1, §27.5
2, 3 Open- vs closed-loop control §27.1
4, 5, 6, 16, 19 Navigation / Kalman state estimation §27.2
7, 8 Guidance, PEG, velocity-to-be-gained §27.3
9, 10, 11, 12, 15 PID control and stability §27.4
13, 14 Sensors → actuators, loop architecture §27.5
17, 20 Autonomy and light-time §27.6