Instructors:
Niklas Fiedler
Displayed in timetable as:
IR - VL
Language of instruction:
English
Min. | Max. participants:
- | 40
Comments/contents:
This lecture covers the most important topics in the field of intelligent robotics.
These include, among others, sensor technology and data processing, interpretation of sensor data and state estimation, architectures for robot control, systems for action planning, and algorithms for continuous control.
The lecture proceeds by presenting sensors for different modalities, their properties and measurement principles, and gives exemplary applications for each of them.
The presented applications illustrate how robotic systems can profit from sensor reading in the different modalities and their integration into dynamic environment models.
Furthermore, the lecture provides theoretical background for the development of modern intelligent systems, including probablistic models and symbolic action planning.
The following topics are covered:
- General sensor characteristics and taxonomy
- Motion sensing: motorencoders, resolvers, inertia measurement units
- Tactile sensing: force- and pressure sensing, force-torque sensors
- Optical sensing: cameras, camera models, visual servoing
- Distance sensing: time-of-flight, phase shift, and triangulation for sensing with infrared, ultrasonic or laser sensors
- Algorithms for state estimation and fusion of sensor data (e.g. ICP, Kalman filters, particle filters)
- Perception-Action-Loop
- Action planning
- Deliberative and behavior-based control architectures
Learning objectives:
The participants learn about various methods of physical sensing, get to know the basics of sensor data processing as well as methods for state estimation and action planning and are able to use these in the context of complex robotic systems.
