Question
Q: What's the key feature/distiction between autonomous systems and traditional systems?
A: Autonomous systems can percept unstructured and unknown environment, perform non-predefined Situations and program test.
Q: What's the differences between hardware and software / algorithmic frameworks?
A: Hardware frameworks:
- Central controller
- Sensors
- Drive / Chassis
Software frameworks:
- Perception
- Planning
- Control
Q: What's the main sensors for navigation?
A: LiDAR, radars, ultrasonic sensors, IMU, GNSS
Q: What's the definition of multi-sensor fusion?
A: Complement each other, re-calculate each other to achieve better accuracy, better reliability and also cut overall costs
Q: What's the main classification and tasks of machine learning?
A: Classification:
- Supervised learning
- Unsupervised learning
- Semi-supervised learning
- self-supervised learning
Tasks:
- regression
- classification
- clustering
Q: What's the definition of overfitting and underfitting?
A:
- Underfitting: The model fails to adequately learn the general patterns of the training data, resulting in poor performance on both the training set and test set.
- Overfitting: The model learns the training data too well, including noise and specific details, causing it to perform strongly on the training data but poorly on new, unseen test data. This weakens the model's ability to generalize.