SBIR Phase I: An Interactive 3D Game of Evolutionary Robotics for STEM-based Education

Period of Performance: 01/01/2014 - 12/31/2014


Phase 1 SBIR

Recipient Firm

Xemory LLC
4 Walden Woods
Essex Junction, VT 05452
Principal Investigator, Firm POC


The SBIR project aims is to excite high school and middle school children about the possibility of working in a STEM-related career. According to the National Center for Education Statistics there are 47 million students enrolled in K-12, with 29 million in middle or high school (6-12). Inspiring and engaging these students in STEM is a key component of the President?s ?Educate to Innovate Initiative?. This project will produce a low-cost computer game designed to heighten interest in STEM-related careers among these students by engaging them in game play and interactive design of robotic controllers and robot body-parts while they apply principles of mathematics, biomechanics, neuroscience and computer science. With distribution through direct internet download and through major app stores, this product taps into the commercial gaming industry with revenues of over $14B annually according to the NDP Group. Partnering with local high schools and university researchers, the project will advance our understanding of how to balance the challenge of gaming with the rewards of learning. This SBIR project proposes to deliver a software game product for Science, Technology, Engineering, Mathematics (STEM) education in the area of evolutionary robotics. The project will explore the use of Interactive Evolutionary Algorithms (IEAs) to evolve the behavior of simulated robots. In collaboration with experts in evolutionary robotics and evolutionary algorithms, the project team will measure the effectiveness of using IEAs as compared to non-interactive simple genetic algorithms and direct controller programming. With an IEA the user influences the direction of evolution through user demonstration. This method can significantly reduce evolutionary time through the search space and provides a more hands-on approach to evolutionary robotics, continually engaging the student user during the design and simulation process. The project will also investigate the use of scaffolding to physically support the robot during the evolutionary process. The physical supports are removed over evolutionary time, gradually increasing the difficulty of the task as the robot evolves, until the robot can walk or run unassisted.