Tyler O'Neal, Staff Editor SYSTEMS

Australian researchers use machine learning to design quieter drone propellers

Researchers have published a study revealing their successful approach to designing much quieter propellers.

The Australian research team used machine learning to design their propellers, then 3D printed several of the most promising prototypes for experimental acoustic testing at the Commonwealth Scientific and Industrial Research Organisation's specialized 'echo-free chamber.

Results now published in Aerospace Research Central show the prototypes made around 15dB less noise than commercially available propellers, validating the team's design methodology.

RMIT University aerospace engineer and lead researcher Dr. Abdulghani Mohamed said the impressive results were enabled by two key innovations - the numerical algorithms developed to design the propellers and their consideration of how noise is perceived in the human ear - as part of the testing. Various propeller prototypes designed by the algorithm

"By using our algorithms to iterate through a variety of propeller designs, we were able to optimize for different metrics such as thrust, torque, sound directivity, and much more. We also formulated a new metric, which involves how the human ear perceives sound, and propose to use that in future designs," he said.

"Our method for optimizing design can be applied to small propellers used on drones to much larger ones used for future urban air mobility vehicles - or air taxis - designed to carry human passengers."

The team, which also included Melbourne-based aerospace company XROTOR, explored how various manipulations of propeller blade noise affected how it was perceived by the human ear.

Mohamed said this modulation had the potential to be used as an important design metric for future propellers.

"The modulation of high-frequency noise is a major factor in the human perception of rotor noise. Human ears are more sensitive to certain frequencies than others and our perception of sound also changes as we age," he explained.

"By designing to such metrics, which take into account human perception, we can design less annoying propellers, which one day may actually be pleasant to hear."

XROTOR Managing Director, Geoff Durham, said it was exciting to see prototype testing show the new designs could significantly reduce the sound impact of drones.

"Not only were the designs appreciably quieter to the human ear, but the propellers had a higher thrust profile against standard market propellers at the same throttle signal input," he said. Echo-free test chamber

The RMIT research team also included Dr. Woutijn Baars, Dr. Robert Carrese, Professor Simon Watkins, and Professor Pier Marzocca. The prototypes were 3D printed at RMIT's Advanced Manufacturing Precinct.

RMIT Aerospace Engineering student Liam Bullard was also involved in the project and said having that opportunity while still, an undergraduate was one of the main reasons he chose to study at RMIT.

"As a student, it's great to get to get to work on projects where I can apply knowledge to a real-world industry problem," he said.

The paper, 'Quantifying modulation in the acoustic field of a small-scale rotor using bispectral analysis' is published in Aerospace Research Central (DOI:10.2514/6.2021-0713).

Bitcoin price boom locked-in vast energy consumption

Tyler O'Neal, Staff Editor SYSTEMS

The cryptocurrency market has been abuzz as Bitcoin gains popularity with investors, reaching an all-time high of over $58,000 apiece in February. In a commentary published March 10 in the journal Joule, financial economist Alex de Vries quantifies how the surging Bitcoin price is driving increasing energy consumption, exacerbating the global shortage of chips, and even threatening international safety.

Theoretically, any computer with access to the internet and electricity can "mine" Bitcoin, a process to receive cryptocurrency by solving sophisticated mathematical equations. It is estimated that all miners combined makeover 150 quintillion--that is 18 zeros following 150--attempts every second to solve the equation, according to numbers from January 11, 2021. Computational power and electricity cost become crucial to profiting from Bitcoins.

"If you're a Bitcoin user making transactions, you're not the one directly paying for electricity. It's a bit of a hidden cost from a user perspective," says author Alex de Vries, the founder of Digiconomist (@DigiEconomist), a blog that highlights new digital trends such as cryptocurrency.

The hidden cost goes beyond energy consumption. Based on the Bitcoin price in January, de Vries estimated that the entire Bitcoin network could consume up to 184 TWh per year, close to the amount of energy all data centers consumed globally. The consumed energy also results in 90.2 million metric tons of CO2, compared to the carbon footprint of metropolitan London.

"That's a pretty mind-blowing number," says de Vries. "Those data centers serve the most of global civilization, and then there's Bitcoin, which serves almost no one but still manages to consume about an equal amount of electricity."

The market price of Bitcoin is an incentive for miners to invest in hardware and electricity. As the price rises, more people put in orders to purchase and run the hardware, causing an increase in energy consumption, and vice versa when the price drops. Due to the overwhelming demand, hardware manufacturers have reported that their devices are sold out, and some customers may not receive their orders until later. This suggests that the amount of energy consumption is "locked-in" at the time of purchase.

"The price of Bitcoin can crash by 25%, 30%, and you may still end up at the same energy consumption point because of the lock-in effect," says de Vries. "The whole idea of my article is to translate what the skyrocketing Bitcoin price is going to mean, not just for the environment, but also externalities that go beyond that."

Bitcoin mining rigs' short shelf-life can mean a substantial amount of electronic waste in the coming years. Mining devices also exacerbate the current global chip shortage by competing for the same chips as personal electronics and electric vehicles, which play an essential role in combatting climate change. Countries with inexpensive electricity, such as Iran, can introduce new revenue streams through Bitcoin mining.

"You can do a lot about these problems. Mining facilities are usually centralized. They're pretty easy to target," says de Vries. Policymakers can intervene by raising electricity rates or confiscating mining equipment. Taxing Bitcoin mining device manufacturers or limiting their access to chips are also strategies to consider. Although Bitcoin is a decentralized currency, government agencies can regulate exchange platforms and prevent its trading to influence the value.

De Vries notes that "we are limited to the information that we have today," and he cautions predictions for future trends regarding Bitcoin. "Who knows what will happen in 2024? Maybe everyone is using bitcoin, maybe nobody, maybe everyone forgot about it could also be the case," he says.

Joule, de Vries: "Bitcoin boom: what rising prices mean for the network's energy consumption"
https://www.cell.com/joule/fulltext/S2542-4351(21)00083-0

FAU researchers win NSF CAREER awards

Tyler O'Neal, Staff Editor SYSTEMS

Researchers represent the College of Engineering and Computer Science

Two researchers from Florida Atlantic University's College of Engineering and Computer Science have received the coveted National Science Foundation (NSF) Early Career (CAREER) awards. The CAREER program offers the NSF's most prestigious awards in support of early-career faculty who have the potential to serve as academic role models in research and education and to lead advances in the mission of their department or organization.

The researchers are Xiangnan Zhong, Ph.D., an assistant professor; and Zhen Ni, Ph.D., an assistant professor, both in the Department of Computer and Electrical Engineering and Computer Science, who received the NSF CAREER awards to drive the current artificial intelligence (AI) wave. With these NSF CAREER awards, the researchers will help to close the gap from the current state-of-the-art techniques to the artificial general intelligence that will bring good performance in learning speed, data efficiency, and generalization of the optimization performance.

"Although existing achievements in artificial intelligence and reinforcement learning are exciting, the fundamental research of data aggregation, learning and approximation capability, and the performance generalization during uncertainties is not yet fully developed," said Stella Batalama, Ph.D., dean, College of Engineering and Computer Science. "The prestigious NSF CAREER awards that professors Zhong and Ni have received will help us to close the gap from the current state-of-the-art techniques to the artificial general intelligence that will bring good performance in learning speed, data efficiency, and generalization of the optimization performance. Moreover, the activities resulting from these projects will vigorously contribute to the nation's artificial intelligence workforce development."

Zhong has received a $503,000 NSF CAREER grant to investigate the intelligent learning control to enable cyber-physical systems (CPS) with the capabilities of autonomous learning and generalization to rapidly adapt in unknown situations. The results are expected to transform how agents interact in high-dimensional and heterogeneous environments, and therefore could potentially provide in-depth findings for exploring creativity in frontier AI techniques.

Zhong also will develop cooperative learning strategies to share with extended skills to facilitate exploration and prevent agents from getting confused by the action details. Also, this project will develop self-motivated learning structures to contribute toward the global objectives for team-wide success in a distributed perspective.

"The integration of research and education plans will prepare our future workforce in the fields of cyber-physical systems, AI, learning, and control," said Zhong.

Ni has received a $500,000 NSF CAREER grant for a natural concurrent reinforcement learning framework that has three major advantages over traditional reinforcement learning methods. These include advantages of simultaneously learning multimodal properties of the complex system; structural advantages of using a personalized learning scheme; and implementation advantages of the data-driven sample-efficient design. Within this framework, Ni will design two concurrent reinforcement learning methods to build the learning-in-learning control paradigm. The applications in the smart energy community will support the novel learning framework and theoretical results.

"Beyond the scientific impacts, this research has broader impacts for a wide range of research disciplines including transportation, rehabilitation, and robotics," said Ni. "Furthermore, the integration of research and education activities will positively impact institutions both regionally and nationally."