Download PrepIQ Autonomous Navigation Systems Ultimate Exam and more Exams Technology in PDF only on Docsity!
Systems Ultimate Exam
Question 1. Which autonomy level corresponds to a system that can perform all driving tasks under all conditions without human intervention? A) Level 2 B) Level 3 C) Level 4 D) Level 5 Answer: D Explanation: Level 5 autonomy denotes full self-driving capability in any environment, requiring no human driver at all. Question 2. In the NED (North-East-Down) local tangent plane, which axis points downward? A) X-axis B) Y-axis C) Z-axis D) None of the above Answer: C Explanation: In NED, the Z-axis is defined to point toward the center of the Earth (down). Question 3. Which mathematical representation avoids gimbal lock when describing 3-D orientation? A) Euler angles B) Rotation matrices C) Quaternions
Systems Ultimate Exam
D) Direction cosines Answer: C Explanation: Quaternions provide a singularity-free representation of orientation, unlike Euler angles which suffer from gimbal lock. Question 4. A MEMS gyroscope primarily measures which physical quantity? A) Linear acceleration B) Angular velocity C) Magnetic field strength D) Temperature Answer: B Explanation: Gyroscopes sense angular rate; MEMS gyros use vibrating proof masses to detect rotation. Question 5. In an IMU, the term “bias” refers to: A) Random noise that averages to zero B) A constant offset in sensor output C) Temperature-dependent drift D) A scaling error in the measurement range Answer: B Explanation: Bias is a fixed offset that must be calibrated out to obtain accurate inertial measurements. Question 6. Which GNSS constellation uses the L1-C signal for civilian services?
Systems Ultimate Exam
Explanation: DGPS transmits correction data from a known base station to reduce common-mode errors. Question 9. In a LiDAR system, the range resolution is primarily determined by: A) Laser wavelength B) Pulse repetition frequency C) Detector bandwidth D) Pulse width (duration) Answer: D Explanation: Shorter laser pulse widths provide finer range resolution because distance is derived from pulse travel time. Question 10. FMCW radar determines target velocity using: A) Time-of-flight measurement B) Phase shift between transmitted and received signals C) Doppler frequency shift of the beat signal D) Pulse-compression techniques Answer: C Explanation: FMCW radar mixes transmitted and received chirps; the resulting beat frequency encodes target range and Doppler shift gives velocity. Question 11. Which camera configuration enables depth perception without active illumination? A) Monocular RGB
Systems Ultimate Exam
B) Stereo (two-camera) rig C) Infrared depth sensor D) Time-of-flight camera Answer: B Explanation: Stereo vision computes disparity between two synchronized images to infer depth. Question 12. In ultrasonic ranging, the speed of sound in air is assumed to be approximately: A) 150 m/s B) 340 m/s C) 500 m/s D) 750 m/s Answer: B Explanation: At room temperature, sound travels roughly 340 m/s, which is used to convert time-of-flight to distance. Question 13. Wheel odometry error grows primarily because of: A) Sensor quantization noise B) Wheel slip and uneven terrain C) Magnetic interference on encoders D) GPS signal loss Answer: B Explanation: Slip and terrain irregularities cause the measured wheel rotation to diverge from actual vehicle motion, accumulating drift.
Systems Ultimate Exam
Answer: C Explanation: UKF captures higher-order moments by passing deterministic sigma points through the nonlinear functions, avoiding Jacobian computation. Question 17. In graph-based SLAM, a “loop closure” constraint is added when: A) The robot revisits a previously mapped area B) A new landmark is detected for the first time C) The IMU bias is updated D) The GNSS signal is lost Answer: A Explanation: Recognizing that the robot has returned to a known location creates a constraint that reduces accumulated drift. Question 18. Monte Carlo Localization (MCL) represents the belief distribution using: A) A single Gaussian B) A set of weighted particles C) A deterministic grid D) A Kalman filter covariance matrix Answer: B Explanation: MCL samples many particles, each representing a possible pose, and weights them according to sensor likelihood. Question 19. Dijkstra’s algorithm guarantees the shortest path in a graph because it:
Systems Ultimate Exam
A) Uses a heuristic function B) Expands nodes in order of increasing cost from the start C) Randomly samples the configuration space D) Prunes edges based on obstacle proximity Answer: B Explanation: Dijkstra explores nodes by the lowest cumulative cost, ensuring optimality without a heuristic. Question 20. The admissible heuristic in A* must satisfy which property? A) Overestimate the true cost to goal B) Be exactly equal to the true cost C) Never exceed the actual minimal cost (i.e., be optimistic) D) Depend on the robot’s velocity Answer: C Explanation: An admissible heuristic never overestimates, ensuring A* remains optimal. Question 21. Rapidly-exploring Random Trees (RRT) are particularly suited for: A) Finding the exact optimal path in a grid B) High-dimensional continuous spaces with complex constraints C) Purely discrete graph search D) Solving linear programming problems Answer: B
Systems Ultimate Exam
B) Feasible velocity commands within the robot’s dynamic constraints and obstacle proximity C) Pre-computed motion primitives only D) Purely reactive sonar readings Answer: B Explanation: DWA evaluates admissible velocities in a “window” defined by acceleration limits and selects the one maximizing a cost function while avoiding collisions. Question 25. Artificial Potential Fields (APF) can suffer from: A) Excessive computational load for large maps B) Local minima where the robot gets stuck C) Requirement for high-precision GPS D) Inability to handle dynamic obstacles Answer: B Explanation: APF creates attractive and repulsive forces; certain configurations can create local minima that trap the robot. Question 26. Trajectory rollout in local planning typically involves: A) Solving a set of linear equations offline B) Simulating candidate motion profiles and evaluating cost metrics C) Using only static waypoints from the global planner D) Ignoring vehicle dynamics Answer: B
Systems Ultimate Exam
Explanation: Rollout generates multiple feasible trajectories, scores them (e.g., clearance, smoothness), and selects the best one. Question 27. Finite State Machines (FSM) are best suited for: A) Continuous optimization problems B) Hierarchical behavior selection with well-defined discrete modes C) Probabilistic reasoning under uncertainty D) High-dimensional motion planning Answer: B Explanation: FSMs model systems as a set of discrete states with transition conditions, ideal for mode-based control. Question 28. Behavior Trees improve over FSMs primarily by: A) Providing a probabilistic framework B) Allowing modular, reusable sub-behaviors and a tick-based execution model C) Eliminating the need for sensor inputs D) Guaranteeing optimality in path planning Answer: B Explanation: BTs organize behaviors as a tree of nodes, enabling composability and clearer execution flow. Question 29. A non-holonomic constraint for a car-like robot forbids: A) Forward motion B) Instantaneous lateral (sideways) velocity
Systems Ultimate Exam
Question 32. Model Predictive Control (MPC) differs from PID by: A) Using a fixed gain for all operating points B) Solving an optimization problem over a finite prediction horizon at each step C) Ignoring system constraints D) Being only applicable to linear systems Answer: B Explanation: MPC predicts future behavior, optimizes a cost function while respecting constraints, and implements only the first control action. Question 33. In state-space representation, the matrix A defines: A) Input-to-state mapping B) State-to-output mapping C) State-to-state dynamics (system matrix) D) Control-to-output relationship Answer: C Explanation: The A matrix captures how the current state evolves autonomously (without inputs). Question 34. Feedback linearization aims to: A) Approximate nonlinear dynamics with a linear model using gain scheduling B) Cancel nonlinearities through a change of variables and control law, yielding a linear closed-loop system C) Perform linear regression on sensor data
Systems Ultimate Exam
D) Replace the plant with a purely proportional controller Answer: B Explanation: By algebraically compensating for known nonlinear terms, the system behaves linearly under the new control input. Question 35. Lateral control for lane keeping commonly uses which sensor as primary input? A) Inertial Measurement Unit (IMU) only B) Front-facing camera detecting lane markings C) Ultrasonic side sensors D) GPS heading alone Answer: B Explanation: Vision-based lane detection provides the lateral error needed for steering corrections. Question 36. Adaptive Cruise Control (ACC) primarily regulates which vehicle variable? A) Steering angle B) Lateral position C) Longitudinal speed relative to a lead vehicle D) Battery state of charge Answer: C Explanation: ACC maintains a desired following distance by adjusting throttle and braking to control speed.
Systems Ultimate Exam
Answer: C Explanation: Early fusion merges raw measurements (e.g., IMU + GPS) before higher-level interpretation to exploit complementary information. Question 40. Redundancy management in safety-critical systems aims to: A) Reduce system cost by eliminating duplicate components B) Detect and isolate failed components while maintaining required functionality C) Increase computational load for better performance D) Replace all hardware with software simulations Answer: B Explanation: Redundant subsystems provide fault tolerance; management monitors health and switches to backups when needed. Question 41. ISO 26262 addresses functional safety for which domain? A) Aerospace avionics B) Marine navigation C) Automotive electric/electronic systems D) Industrial robotics Answer: C Explanation: ISO 26262 is the international standard for functional safety of road vehicles. Question 42. DO-178C is a certification standard primarily used in: A) Automotive ECU software development
Systems Ultimate Exam
B) Aerospace software for airborne systems C) Medical device software D) Consumer electronics firmware Answer: B Explanation: DO-178C defines objectives for safety-critical software in avionics. Question 43. Hardware-in-the-Loop (HIL) testing provides: A) Pure software simulation without any real hardware B) Real-time execution of controller code interfaced with simulated plant models C) Only post-deployment field testing D) Manual inspection of printed circuit boards Answer: B Explanation: HIL runs the actual embedded code while the plant dynamics are modeled in real time, enabling early validation. Question 44. The Gazebo simulator is commonly used for: A) High-fidelity aerodynamic CFD analysis B) Real-time 3-D robotics simulation with physics and sensor plugins C) Satellite orbit propagation D) Audio signal processing Answer: B Explanation: Gazebo offers physics-based environments and virtual sensors for robotics research.
Systems Ultimate Exam
Answer: B Explanation: Hard-iron effects are fixed offsets that shift the measured magnetic field, requiring calibration. Question 48. Ring Laser Gyro (RLG) advantages over MEMS gyros include: A) Lower cost and smaller size B) Higher precision and lower drift over time C) Ability to measure linear acceleration directly D) Immunity to temperature changes Answer: B Explanation: RLGs provide very low bias instability, making them suitable for high-accuracy navigation. Question 49. In GNSS, the term “Dilution of Precision (DOP)” quantifies: A) Signal strength attenuation due to atmosphere B) The geometric quality of satellite configuration affecting position error C) The number of visible satellites D) The clock synchronization error Answer: B Explanation: DOP is a scalar factor derived from satellite geometry; lower DOP indicates better positional accuracy. Question 50. A “carrier-phase” GNSS measurement is used primarily for: A) Coarse positioning with meter-level accuracy
Systems Ultimate Exam
B) High-precision positioning (centimeter-level) in RTK applications C) Determining satellite health status D) Measuring ionospheric delay directly Answer: B Explanation: Carrier-phase tracking resolves the phase of the carrier wave, enabling sub-centimeter accuracy when integer ambiguities are resolved. Question 51. In a SLAM system, “data association” refers to: A) Matching sensor measurements to previously mapped landmarks B) Synchronizing timestamps across sensors C) Converting raw data into binary format D) Broadcasting pose estimates over the network Answer: A Explanation: Data association determines which observed features correspond to existing map elements, crucial for consistent mapping. Question 52. The Jacobian matrix in EKF is used to: A) Transform sensor data into the robot frame B) Linearize the nonlinear motion and measurement models around the current estimate C) Compute the inverse of the covariance matrix D) Generate random samples for particle filters Answer: B Explanation: EKF propagates mean and covariance through linear approximations given by Jacobians of the nonlinear functions.
