TurtleTalk

Inspiration

We wanted to make TurtleBot more accessible, especially for those without extensive computer skills. We also are deeply interested in voice control technology and robotics so we wanted to integrate the two passions.

What it does

Interpret commands to Amazon's Alexa and send them through a personal server (Pluto) to ROS on a TurtleBot computer and eventually to the TurtleBot itself

How we built it

We developed an Amazon skill. To integrate this with ROS we attempted to create an AWS lambda but determined that a personal server would be more efficient and simple. We then connected the Amazon skill to the personal server which transmitted key Distance, Direction, and Speed data to ROS. From ROS, we attempted to parse this data into actionable commands.

Challenges we ran into

We struggled to work with AWS lambda since it lacks an IP address and so makes sending data difficult. We solved this problem by developing our own server which we used instead. We struggled to connect networks because of local firewalls but eventually found loopholes around this problem. We did not have sufficient time to develop full integration between Alexa on Echo and our skill and between ROS data and the TurtleBot.

Accomplishments that we're proud of

Successfully developing an Amazon skill. Problem solving our difficulties with AWS lambda by using our own server. Navigating a daisy chain of servers and firewalls to successfully send data from Alexa test simulations to ROS.

What we learned

How to manipulate servers to successfully send relevant data. How to work together in shifts so that we collectively receive enough sleep to continue coding productively and intelligently.

What's next for TurtleTalk

Consistently receiving commands in TurtleBot from ROS.

Built With

• alexa-skills-sdk
• amazon-alexa
• python
• ros