Lesson 1GPS/IMU an time synchronization: RTK, PPP options, IMU drift characteristics, timestampin an synchronization protocolsDis section introduce GPS, IMU, an timin requirements. It compare RTK an PPP, describe IMU drift characteristics, an explain timestampin an synchronization protocols needed fi precise, low-latency sensor fusion an localization.
GNSS accuracy an availability limitsRTK an PPP correction strategiesIMU bias, noise, an drift modelsTime bases an timestamp policiesPPS, PTP, an IEEE 1588 usageClock monitorin an fault handlinLesson 2Perception stack components: detection, classification, trackin, lane model estimation, gap acceptance estimationDis section decompose di perception stack into detection, classification, trackin, an lane modelin. It also cover gap acceptance estimation an how dese components interact fi support safe lane keepin an maneuver decisions.
Object detection an region proposalsObject classification an attributesMulti-object trackin an ID managementLane model estimation an qualityGap acceptance an TTC estimationInterfaces to plannin an controlLesson 3Sensor roles by function: segmentation of responsibilities fi lane keepin, object detection/trackin, an localizationDis section assign functional roles to each sensor type. Learners see how radar, lidar, cameras, an GNSS/IMU share responsibilities fi lane keepin, object detection an trackin, an localization in a balanced, fault-tolerant design.
Lane keepin sensin responsibilitiesObject detection an confirmation rolesLongitudinal an lateral trackin dutiesLocalization an map alignment rolesRedundancy an graceful degradationRole allocation fi highway pilotLesson 4Calibration, extrinsics, an online self-checks: calibration verification, boresight checks, an integrity monitorinDis section focus pon calibration an integrity monitorin. It cover extrinsic an intrinsic calibration, boresight checks, online self-checks, an health metrics dat detect misalignment or sensor faults before dem degrade safety.
Intrinsic calibration of cameras an lidarExtrinsic calibration between sensorsBoresight checks fi radar an camerasOnline self-checks an residual testsHealth metrics an fault thresholdsRecalibration triggers an workflowsLesson 5Typical automotive sensor specs: front radar (ranges, resolution, update rate, field of view)Dis section review front radar specifications an dem impact pon design. It cover range, range an velocity resolution, update rate, field of view, an how dese parameters affect highway cut-in detection, trackin stability, an safety margins.
Maximum an minimum detection rangeRange, angle, an velocity resolutionUpdate rate an trackin latencyHorizontal an vertical field of viewMulti-path, clutter, an interferenceHighway pilot radar performance needsLesson 6Typical automotive camera specs: resolution, frame rate, dynamic range, lens FOV, mountin an calibration needsDis section cover camera specifications relevant to autonomous drivin. It address resolution, frame rate, dynamic range, lens field of view, an mountin an calibration needs, linkin each to lane detection, object recognition, an fusion.
Image resolution an pixel sizeFrame rate an exposure controlDynamic range an HDR techniquesLens FOV an distortion profilesMountin rigidity an placementIntrinsic an extrinsic calibrationLesson 7Typical automotive lidar specs: range, angular resolution, point rate, weather performance, mountin considerationsDis section explain automotive lidar specifications an tradeoffs. Learners examine range, angular resolution, point rate, weather an contamination performance, an mountin constraints dat influence coverage, occlusions, an fusion design.
Detection range an reflectivity limitsHorizontal an vertical angular resolutionPoint rate, scan pattern, an densityRain, fog, an dust performanceVibration, height, an occlusion issuesCleanin, heatin, an contaminationLesson 8Sensor fusion architectures: low-, mid-, high-level fusion tradeoffs an recommended approach fi highway pilotDis section survey sensor fusion architectures an tradeoffs. It contrast low-, mid-, an high-level fusion, den motivate a recommended mid-level approach fi highway pilot, focusin pon robustness, latency, an implementation complexity.
Low-level fusion an raw data sharinMid-level fusion wid object listsHigh-level fusion of decisionsLatency, bandwidth, an compute costsFailure isolation an redundancyHighway pilot fusion reference designLesson 9HD map data attributes: lane-level geometry, speed limits, merge tags, lane connectivity, confidence an versioninDis section detail HD map lane geometry, attributes, an metadata. Learners see how speed limits, lane connectivity, merge tags, confidence, an versionin support plannin, localization, an safe behavior in changin road networks.
Lane centerlines an boundariesLane-level speed limits an rulesMerge, split, an turn lane tagginLane connectivity graphs an topologyConfidence scores an freshness flagsMap versionin an change management
