By using a drone to complete this portion of the competition, the RovoNav team can avoid terrain-related traversability challenges and reduce the weight of the rover. The mechanical team can simplify the rover’s design and sensor load, and instead focus on designing and manufacturing other complex structures such as the arm, end effector, chassis, and soil sampling subsystem. The software team chose a prefabricated drone (the VOXL m500) and is building the flight software using PX4 and ROS2. 

To ensure a safe landing, the team will use a map interface to select and feed the drone “good” coordinates in the vicinity of the autonav gates. The gates progressively increase in traversability difficulty from the first to the last (eighth) gate, and the drone will take the unusual strategy of navigating from gate 8 to 1, essentially completing the course backwards. The software team plans to use this strategy to have the gates with the most treacherous surrounding terrain completed by the drone. Since the drone cannot complete the course on one battery charge after accounting for payload, the rover will then complete the course by moving through the remaining easy-to-access gates in the normal order, from gate 1 to wherever the drone stopped. In this unique and piecewise way, the team hopes to complete the course successfully and surpass other teams in the competition.

A small contingent of RoboJackets set up a table with posters and example robots to let the wider community know what kinds of robotics projects are out there, and also were introduced during a timeout.

Zach Slaton, President, was present and said “Tabling at the basketball game gave us some great publicity to people we wouldn’t normally meet, and we were especially popular with kids. We gave out a lot of stickers, and we hope we inspired some future RoboJackets!”

The 12-pound Hockii is a type of combat robot known as a “meltybrain”, which uses its drive to spin and hit opponents with the entire body [read more about the previous iteration here]. After last year’s competition, the team noticed that the weapon tooth was a major weakness. Additionally, the fact that it was integrated with the frame meant it was impossible to replace. Thus the team aimed to make a tooth that was both stronger and replaceable. 

Mounting the tooth to the frame introduces additional challenges. An improperly-designed tooth or joint would result in the fasteners breaking and the robot ending up toothless. Hockii’s new tooth slots into the frame to help direct moment loads into the shell, while two large bolts keep it in place.

Team member Martha Leach took the tooth design to the next step. Through her internship at the Mayo Clinic Division of Engineering, she had the opportunity to research and utilize computational design and additive manufacturing tools. She led the team to develop experimental teeth which leverage two unique techniques: topology optimization and generative design. 

Topology optimization is a computational technique where a model is iteratively modified to minimize an objective while still meeting a target constraint. In this case, the model was optimized to reduce material while still being able to withstand the load of impact. 

Generative design is a newer technique where a designer starts by inputting constraints and features of a part. The software then explores the design space to generate multiple possible parts which satisfy the prescribed constraints.

These two teeth were then printed out of titanium using the Selective Laser Melting (SLM) printing process. The teeth are pictured below. 

The team behind Hockii is hard at work preparing all robot systems for the March NHRL competition and hopes to put these new teeth to the test against some of the top 12lb combat robots in the nation. 

When the team first began evGP, a few hardware components were required as part of the rules to be standardized across the robots. Some of those components included a specific Top Kart frame, batteries, and motor component combination. This was originally to give equal footing to new competitors and allowed for the teams to focus on a reduced quantity of components and systems not specified by the rules.

For the drive system, all of the teams were required to use a brushed motor (ME0708) with a golf-cart motor controller (SR48300) and an optional reversing contractor (SW202). Using current sensors, encoders, and practical application of Georgia Tech controls classes, the team was able to compete in two evGP competitions with the drive system. However, there were some control bugs with the golf-cart motor controller. This caused “interesting” events like the robot unexpectedly driving forwards / backwards and low speed control resulting in the car starting and stopping constantly. Additionally, talks in the competition planning had resulted in the expansion of possible drive systems, specifically using brushless motors. 

Brushed style motors, as was originally on the robot, were some of the most widely used motors for many years in industrial automation. They were very easy to actuate with only a voltage potential being applied across the two wires and are still very often found in cheap toys because of the simplicity. Brushless motors, in comparison, have three wires that need specific timing to turn on and off in a pattern, with this added complexity limiting adoption in smaller scale / hobbyist applications. Brushless motors have since exploded in popularity and availability due to more efficient energy/weight ratio and longer lifetime compared to brushed motors.

This combination of factors has led the team to switch to a new brushless motor (Golden Motor HPM5000B) and motor controller (VESC 75/300R3) for actuating the drive system. The new VESC controller includes the capabilities to measure current and voltage along with customizable filters, proving much more sophisticated than the original motor controller and allowing for more tightly integrated speed control. The new system also allows for the removal of the reversing contactor, which was heavy and generated lots of heat in the closed enclosure, while maintaining the reversing ability. RPM control has been proven with the motor and controller, with additional development required for full robot drive control.