BHR 09

This seminar will discuss the interactive aspects of humanoid robots. We will cover algorithms that enable intelligent behavior around humans through physical interaction, conversation and other forms of negotiation with other beings in the environment. This will include existing and future applications of the robots we are building.

We continue our discussion of whole body motion. Here we would like to focus more on specific algorithms that enable the humanoid robot to combine its degrees of freedom to generate greater capabilities for executing tasks than would be expected of separating balance, manipulation, and grasping strategies.

We will focus on how a humanoid robot not only uses its wheels or its arms to accomplish tasks. In fact, it uses its entire body consisting of all these joints. By combining the capabilities of its actuators and sensors we can accomplish more tasks. However, we also need algorithms that can handle this complexity.

The two topics presented will relate to force control and the movement of multiple links to achieve this goal.

This seminar will discuss the humanoid ability to grasp objects. How does a robot make decisions on the approach and positioning of its hands with respect to an object in order to grasp it. Which factors about the object shape and robot's hand affect this decision?

This seminar will focus on a more in-depth analysis of how learning can be applied to the control of humanoid robots in specific situations. Either this seminar will deal with learning sensory-motor control of the robot or with its ability to handle complex objects in a scene.

We will go more in depth into specific types of motion planning. What sort of planning can be applied in unusual situations - for example moving a fridge or crawling under a table. How does a robot decide which actions to take?

We will discuss how humanoid robots can learn from both people and their experience in order to get a better understanding of the world. How can learned knowledge be represented and applied to controlling the way a robot performs tasks?

This seminar will discuss general planning algorithms and their applications to humanoid robots. How do humanoid robots make decisions about which joints to move and when to move them? What processes are required for efficient planning in high dimensions? The presentation also briefly discusses on the footstep planning for humanoid robots.

This seminar will cover applications of vision to humanoid robotics. We will go in-depth into specific aspects of vision algorithms that are relevant to humanoid robot platforms.

This seminar will discuss bipedal locomotion as applied to humanoid robots. How do humanoid robots maintain balance while walking? What can we learn about balance from their strategies?

This seminar discusses both the proprio and extero sensing capabilities of robot platforms; also commenting upon the interoceptive sensing capabilities. A broad overview of tactile, proximity and navigational sensing with RFID is presented. Several types of encoders are shown as well. A series of vision algorithms are also presented in the second half of the presentation including SIFT, a description of stereo vision and finally SLAM.

This seminar will discuss robot platforms that balance forces, particularly gravity forces. It will cover the basics of dynamic vs. static balance and explain existing strategies for controlling a balancing platform. Examples may include the segway as well as other wheeled platforms.


For the best explanation of the dynamics of balancing robots please see the BallBot paper as it gives names to the basic components.

Will begin with a working definition of a humanoid robot and use this to explore several humanoid platforms in current development. Discussion will include explanation of successful humanoid technologies and how they may be integrated into future systems. There will also be a brief demonstration of a working humanoid performing basic real-world functions. Attached are several papers documenting work on systems used on a variety of humanoids - these are mainly concerned with motion control.