Publications

Autonomous robots are becoming more common in different sectors, such as agriculture. The use of robots in harvesting allows to cut down human labor costs and makes performing of tasks more efficient. Multiple smaller harvesting robots can also cover more complex harvesting paths. Also, when using an electric robotic fleet instead of large traditional harvesting machines, transportation costs and environmental load are reduced. Thanks to its ability to grow rapidly, reed harvesting can be an effective solution to gain biomass for different applications and to partially replace raw materials that are traditionally made from wood. This study focused on the different challenges in development of a robotic fleet harvesting system and introduced a concept including recognition of the reed field visually and harvesting path planning. Therefore, this study guided towards further investigation and developing of an autonomous robot-based harvesting system that can be applied in agriculture.

Fluid power systems can be enormously complex. The chiefest complexity is in their control components due to the deep knowledge required for effective development, the need to interface with myriad heterogeneous components, and the opacity of its operation. In the IT sector, a tech stack is a known discretization of function within a system, typically used to categorize what software tools do. This enables disparate parts of a system to be developed separately. Inspired by IT tech stacks, this paper presents a model, the mechatronic tech stack, to mitigate this complexity by siloing control system behavior into discrete functions and classifying what tools can be used to meet them. The model is demonstrated on three case studies of diverse application. The theory is promising, but there are challenges in defining the most useful abstractions and interfaces.

Mechatronic systems can be enormously complex. The chiefest complexity is in their control components due to the deep knowledge required for effective development, the need to interface with myriad heterogeneous components, and the opacity of its operation. In the IT sector, a tech stack is a known discretization of function within a system, typically used to categorize types of software tools and the functions they perform. This enables disparate parts of a system to be developed separately with the confidence that other components will or should provide expected interfaces. Inspired by IT tech stacks, this thesis presents a model, the mechatronic tech stack, to mitigate this complexity by siloing control system behavior into discrete functions, classifying what tools can be used to meet them, and how development can be guided by this theory. The model is demonstrated on four case studies related to control upgrades on a fluid power test bench, development of soft robots, the development of a pneumatic muscle Stewart platform, and controllers for a solid mechanics lab. The theory is promising and there is a clear direction to progress, but there are challenges in defining the most useful abstractions and interfaces.

This paper describes work on a concept for a probe that would be capable of km-scale deep subsurface ice penetration on Europa or other Ocean World. Penetrating deep into or through the ice cap over the liquid ocean is the best way to establish if life has evolved there. A central thesis of this work is that we must start by addressing the Planetary Protection constraints, and not to try to add them on at the end. Specifically, all hardware in the probe would be designed to survive heat sterilization at 500C for extended periods, as required to meet the COSPAR 1-in-10,000 probability per mission of biological contamination of the ocean. The baseline concept features a heat source containing plutonium-238 encased within a stainless Dewar so that the heat is not lost by conduction into the ice. A circular saw blade sticks out through a slot in a hemispherical turret dome at the bottom of the Dewar such that the blade cuts the ice as it spins. The turret also rotates slowly to cause the saw blade to make a hemispherical cut in the ice. The ice chips would be thrown up through the slit into the turret and would be melted by the heat source. The meltwater drains into sumps on either side of the sawblade, from which the meltwater would be pumped out to the rear of the probe.

The main body of the probe contains a spool of aluminum tubing that would be dispensed from within the probe all the way back to the lander. This tubing is nominally 1-3 mm in outside dimension with integral insulated electrical wires around the center hole. This tube would pneumatically transport small (mm-scale) single-use canisters containing meltwater samples from the probe to the surface for analysis. Surface analysis allows use of instruments that can be neither miniaturized nor sterilized sufficiently to go down-hole. Dry inert gas would be used to push the canister down from the lander and back up. The dry gas would be re-compressed and re-used. During a portion of the cruise to the outer solar system, the heat source would heat the entire probe, including the coiled tubing inside as well as all the canisters and inert gas, to 500C to destroy any lingering organisms and to decompose any complex organic molecules.

The paper describes analysis, design, and preliminary testing, as well as plans for building a prototype and testing it in an "ice treadmill" where a plug of ice is created inside a vertical LN2 cold jacket, pushed up by water pumped below so that the ice plug rises at the same rate that the probe penetrates the ice.