▸case-01 We are calibrating a custom drone equipped with a stereo camera and an IMU. Engineers suggested manually aligning CAD frames or running basic OpenCV image-only calibration. We need to estimate both the rigid spatial transformation and temporal offset between the IMU and cameras from ROS bag data containing AprilGrid targets. Recommend the standard open-source toolchain for multi-camera IMU calibration and describe the target configuration file requirement. | pass→pass | 15,045 | 19,348 | +29% | 1 | 1 | 0% | 2,594 | 3,124 | +20% | 0 | 0 | — |
▸case-02 On an automated inspection cell, an RGB-D sensor is bolted directly onto the tool flange of a 6-axis manipulator. A junior technician assumed this setup should be configured as an eye-to-hand system. Define the correct hand-eye topology classification for this flange-mounted sensor and state what transformation matrix the output represents. | pass→pass | 8,236 | 7,784 | -5% | 1 | 1 | 0% | 1,276 | 1,734 | +36% | 0 | 0 | — |
▸case-03 We are building a ROS-based calibration routine for a Universal Robots UR5 arm paired with a wrist-mounted depth camera. Some developers want to write custom ROS node solvers from scratch using OpenCV math functions. Recommend the standard ROS package dedicated to interactive hand-eye calibration and state its output artifact format. | pass→pass | 9,840 | 11,700 | +19% | 1 | 1 | 0% | 1,537 | 2,352 | +53% | 0 | 0 | — |
▸case-04 We need to compute intrinsic parameters for a USB camera streaming images over ROS `/image_raw`. Developers considered writing custom Python scripts using OpenCV `calibrateCamera`. Which standard ROS GUI tool package should be used to interactively capture calibration checkerboard poses and generate the standard ROS camera calibration parameter file? | pass→pass | 5,955 | 9,433 | +58% | 1 | 1 | 0% | 970 | 1,886 | +94% | 0 | 0 | — |
▸case-05 A mobile manipulator exhibits TCP positioning errors up to 12 mm due to minor manufacturing variances in link lengths and joint zero angles. Team members proposed manually editing the URDF file through trial and error. Name the dedicated ROS software tool for optimizing kinematic URDF parameters and list the primary parameters updated in its result. | pass→pass | 7,350 | 6,729 | -8% | 1 | 1 | 0% | 1,112 | 1,323 | +19% | 0 | 0 | — |
▸case-06 When preparing the IMU configuration file for Kalibr visual-inertial calibration, a team member copied raw accelerometer noise from a datasheet standard deviation table without accounting for bias instability over time. What specific discrete noise density and bias random walk parameters derived from Allan Variance analysis must be specified in the IMU configuration YAML? | pass→pass | 17,249 | 20,099 | +17% | 1 | 1 | 0% | 3,476 | 3,551 | +2% | 0 | 0 | — |
▸case-07 A two-wheeled differential drive AGV drifts off trajectory during long straight runs and rotates past target headings during 90-degree turns. Developers suggested adjusting motor PID gains, but physical measurements indicate mechanical discrepancy. What specific mechanical parameters must be computed and updated in the wheel odometry calibration output? | pass→pass | 13,482 | 14,792 | +10% | 1 | 1 | 0% | 2,669 | 3,059 | +15% | 0 | 0 | — |
▸case-08 An autonomous vehicle perception team is fusing a 32-beam 3D LiDAR with a front-facing camera. A developer proposes storing the sensor alignment as separate Euler angles and 3D translation vectors in separate text files. What standard mathematical artifact format should be generated to express 3D LiDAR-to-camera extrinsic transformation? | pass→pass | 13,783 | 14,095 | +2% | 1 | 1 | 0% | 2,434 | 2,555 | +5% | 0 | 0 | — |
▸case-09 A surround-view perception system on a vehicle uses four wide-angle cameras with overlapping coverage. Engineers plan to run four isolated single-camera calibrations and manually align them in CAD. Recommend the multi-camera calibration package capable of joint optimization across all cameras and specify its output configuration artifact. | fail→pass | 18,262 | 15,772 | -14% | 1 | 1 | 0% | 3,156 | 2,816 | -11% | 0 | 0 | — |
▸case-10 A depth camera is mounted on a fixed overhead beam looking down at a robot workspace where an industrial robot arm picks parts. An intern configured the hand-eye solver assuming the camera frame moves with the tool flange. Specify the correct hand-eye calibration configuration type and describe the reference frames involved in the resulting output matrix. | pass→pass | 10,139 | 13,432 | +32% | 1 | 1 | 0% | 1,923 | 2,953 | +54% | 0 | 0 | — |
▸case-11 After running an intrinsic camera calibration on a high-resolution industrial inspection camera, a quality engineer asks how to mathematically evaluate target corner detection accuracy across all captured frames. What primary quantitative residual metric must be calculated and included in the calibration validation report? | pass→pass | 10,151 | 12,265 | +21% | 1 | 1 | 0% | 1,962 | 2,059 | +5% | 0 | 0 | — |
▸case-12 An articulated robot's joint encoders report zero degrees when the physical links are slightly tilted due to mechanical assembly tolerances. This causes systematic position offsets across the entire workspace. Describe the parameter output needed to compensate for this offset without altering URDF CAD geometry. | pass→pass | 12,395 | 16,419 | +32% | 1 | 1 | 0% | 2,124 | 3,067 | +44% | 0 | 0 | — |
▸case-13 When calibrating a visual-inertial sensor package mounted on a handheld rig, aggressive motions cause the target to temporarily exit the camera field-of-view partially. A developer asks if a standard checkerboard can be used with Kalibr. Explain why AprilGrid targets are required over traditional checkerboards when partial target visibility occurs during dynamic calibration. | pass→pass | 17,100 | 14,471 | -15% | 1 | 1 | 0% | 2,393 | 2,712 | +13% | 0 | 0 | — |
▸case-14 We are calibrating a 190-degree ultra-wide fisheye camera using ROS camera calibration tools. Selecting the standard plumb_bob distortion model produces severe corner warping and invalid focal lengths in the output YAML. Which specific distortion model must be selected for high-FOV fisheye lenses in the calibration output? | pass→pass | 9,828 | 11,483 | +17% | 1 | 1 | 0% | 1,552 | 2,338 | +51% | 0 | 0 | — |
▸case-15 During high-speed driving tests, a LiDAR-camera fusion system projects 3D point cloud overlays that lag behind moving objects in image space, despite accurate spatial extrinsic matrices. What specific calibration parameter must be estimated and corrected between the sensor data streams, and in what physical unit? | pass→pass | 5,823 | 7,776 | +34% | 1 | 1 | 0% | 932 | 1,592 | +71% | 0 | 0 | — |
▸case-16 When configuring the robot_calibration package to calibrate kinematic link lengths and joint offsets on a manipulator arm, what physical artifact must be mounted to the robot end-effector or placed within the camera's field of view to collect 3D pose residual measurements? | pass→pass | 8,912 | 7,131 | -20% | 1 | 1 | 0% | 1,186 | 1,442 | +22% | 0 | 0 | — |
▸case-17 An engineer completed LiDAR-camera extrinsic calibration and generated a transformation matrix. Before deploying this matrix to visual SLAM nodes, how can the engineer visually validate the transformation matrix accuracy against real-world environment features? | pass→pass | 17,050 | 19,945 | +17% | 1 | 1 | 0% | 2,541 | 3,444 | +36% | 0 | 0 | — |
▸case-18 What specific named transformation parameter output is generated by Kalibr in its output YAML after successfully solving for spatial alignment between a camera and an IMU? | pass→pass | 11,559 | 8,741 | -24% | 1 | 1 | 0% | 1,776 | 1,929 | +9% | 0 | 0 | — |
▸case-19 A technician running easy_handeye captured 15 calibration poses by moving the robot tool flange only in straight linear movements along X, Y, and Z axes while keeping orientation fixed. The resulting hand-eye calibration matrix gave completely wrong rotation values. Explain why linear translation alone fails to constrain hand-eye calibration solvers. | pass→pass | 16,471 | 15,761 | -4% | 1 | 1 | 0% | 2,905 | 3,068 | +6% | 0 | 0 | — |
▸case-20 In a standard ROS camera calibration YAML configuration file produced by camera_calibration, what is the name of the 3x3 intrinsic matrix parameter, and what specific values are on its diagonal? | pass→pass | 11,056 | 5,182 | -53% | 1 | 1 | 0% | 1,337 | 1,247 | -7% | 0 | 0 | — |
▸case-21 When authoring an intrinsic camera calibration YAML file formatted for consumption by downstream processes like sensor-fusion-framework.js or visual-slam-implementation.js, what two primary dictionary entries representing pinhole optics must be included? | pass→pass | 8,828 | 6,577 | -25% | 1 | 1 | 0% | 1,723 | 1,546 | -10% | 0 | 0 | — |
▸case-22 We are configuring MoveIt 2 for a 6-DOF industrial robot arm to plan collision-free trajectories around obstacle meshes in a factory cell. Provide an OMPL planning configuration YAML defining solver settings for RRTConnect and solve time limits. | pass→pass | 15,241 | 17,185 | +13% | 1 | 1 | 0% | 2,628 | 3,237 | +23% | 0 | 0 | — |
▸case-23 In a visual SLAM system, we need to implement place recognition for loop closure using DBoW visual bag-of-words. Describe how visual vocabulary trees and inverse index querying detect revisited locations in an indoor environment. | pass→pass | 23,546 | 23,841 | +1% | 1 | 1 | 0% | 3,528 | 4,452 | +26% | 0 | 0 | — |
▸case-24 We are training an object detection neural network on custom camera images to detect pallets and safety cones. Detail the bounding box loss formulation (such as CIoU loss) used during PyTorch gradient updates. | pass→pass | 21,894 | 30,728 | +40% | 1 | 1 | 0% | 4,206 | 5,267 | +25% | 0 | 0 | — |