Intel's America's Greatest Makers - Season 2 Proposal
Below are two ideas pitched to the producers to Intel's America's Greates Makers
(AGM) Season 2. The two ideas are titled L-CAS: Local-frame Collision Avoidance System and GESTR: Generally Expressive Telepresence Robot
1) L-CAS: Local-frame Collision Avoidance System
L-CAS is a local-frame localization system that allows for increased situational awareness in homogeneous and heterogeneous mixtures of autonomous systems and humans. Each autonomous agent (robot and/or human) can carry an L-CAS unit. The relative distance between each L-CAS unit is determined using a muli-path resistant radio frequency ranging system. These distances are then shared between all L-CAS units. The relative formation of the group is determined by the using the distance from each L-CAS and implementing two and/or three-dimensional trilateration. Trilateration results in the localization of each L-CAS unit in the local frame. The muli-path resistant nature of the ranging system means that this system can be used in non-line-of-sight circumstances (e.g. separated by walls, trees, etc.) while retaining formation information. This information increases situational awareness of all autonomous units.
Use Cases:
First responder situational awareness - Firemen, SWAT, etc.
Autonomous Formation Keeping for swarm and collaborative robots
Autonomous Collision Avoidance
Collaborative SLAM Map Building
Increase Situational Awareness for robot-robot, human-robot, and robot-robot teams
GPS Coverage Expansion
Targeted Intel items used:
Intel Curie - sensors, filter computation, and communications
A next-generation Intel Atom processor-based platform - trilateration computation - user interface (via graphics and gesture recognition using the ZR300 and a touch interface via the high speed I/O)
RealSense ZR300 for user input via gesture recognition as stated above.
2) GESTR: Generally Expressive Telepresence Robot:
GESTR is a telepresence robot that allows a user to interact with a foreign environment and most importantly allows the people in the foreign environment to better interact with the user. This is done by the addition of articulated arms and neck to the telepresence robot. The arms will not be for manipulation but rather for displaying of gestures. The RealSense ZR300 will be used to detect the gestures on the users side. Those gestures will be transported over the cloud to the robot. A second RealSense ZR300 will be used on the robot to ensure it does not hit any humans and/or objects when gesticulating.
Multiple different sizes of robot will be made for different types of users. The adult size robot will be used for live presentations and class lectures. The child size robot will be used for meetings. While the adult size robot will use SLAM via the ZR300 while wondering the hallways and classrooms, the child size robot will use SLAM via the ZR300 to navigate the conference table and not hit coffee cups and not fall off the table.
Note: This device will help me teach remotely while I am abroad on conference and/or doing foreign research (something I do often and thus is much needed).
Published Work: D. M. Lofaro, M. Bula, P. Early, E. Eide and M. Javid, "ARCHR - Apparatus for Remote Control of Humanoid Robots," Humanoid Robots (Humanoids), 2015 IEEE-RAS 15th International Conference on, Seoul, 2015, pp. 229-236.
Published Work: D. M. Lofaro, A. Asokan and E. M. Roderick, "Feasibility of cloud enabled humanoid robots: Development of low latency geographically adjacent real-time cloud control," Humanoid Robots (Humanoids), 2015 IEEE-RAS 15th International Conference on, Seoul, 2015, pp. 519-526.
Published Work: D. M. Lofaro and A. Asokan, "Low latency bounty hunting and geographically adjacent server configuration for real-time cloud control," 2016 IEEE International Conference on Robotics and Automation (ICRA), Stockholm, 2016, pp. 5277-5282.
Video Example: https://www.youtube.com/watch?v=OHAWwdHGb60