» Issue 29: Spring 2001

The flying summer in Delaware

Yuen Yoong Leong (Elec 1997) reports on an idyllic summer working in the United States thanks to a Jessel Rosen Research Award

Initiated by Professor Sinclair Goodlad of Imperial College and Dr Joan Bennett of University of Delaware, the IC-UD Exchange Programme brings six Imperial College students across the Atlantic Ocean each year to undertake eight weeks of undergraduate research in the serene and classic university town of Newark, Delaware. I was one of the participants in summer 2000.

Precision farming is a means of growing crops by making more efficient use of inputs such as fertilisers and pesticides. The expected benefits include increased farm profits with less possibility of environmental damage from agricultural operations. I began to take an interest in it after discovering that this method of farming will be strongly featured in my sponsoring company's research and development in the future. I was absolutely delighted when I found out that there was an ongoing project Non-linear Control with Applications in Precision Farming Machines at the University of Delaware, directed by Professors J Glancey and JQ Sun. I applied to work for them and was acccpted.

Development of fly-by-wire

The world of agriculture is constantly changing. New Holland is one of the major agricultural machine manufacturers that aims to help customers to capitalise on new opportunities, shifting market conditions and emerging technologies. One of the projects being developed at their North American Technical Centre is The Demeter Project. The objective is to develop a fly-by-wire windrower -a machine designed for large-scale farmers to cut and condition forage crops at high speed. In a fly-by-wire digital control system, there is no direct link between the driver-operated control stick and pedals, and the control surfaces. All measured signals, including the driver inputs, are processed by an onboard control computer, which computes the desired control surface deflections. A fly-by-wire system is characterised by the absence of any hydraulic or mechanical components and has a higher safety level than mechanical control systems. Once the windrower is made fly-bywire, the driving of the machine can be automated by using software, Global Positioning System and wireless communications technologies. In order for the machine to operate safely, a closedloop control on the wheel speed is thus needed.

Early in 2000, Professors J Glancey and JQ Sun were invited by New Holland to work on this project. Their role was to design and implement a closed-loop controller for the windrower wheel speed that would provide the best control of the machine and a quick system response time. As their undergraduate research student, my tasks were to perform system identification and controller design.

Yuen Yoong Leong working on the windrower in New Holland

Machine testing amongst the Amish

My initial responsibility was to assist another student in carrying out step response experiments on the windrower to record its behaviour. This task sent the two of us on several enjoyable trips along the winding country roads of Pennsylvania -cutting through vast cornfields, whizzing past Amish buggies and the colourful clothes lines of the Amish people -to the New Holland North American Technical Centre.

Wheel speed as a function of current was a crucial relationship that we tried to determine, because the input current to the motors was the only parameter that we had direct control over. Besides the input current, the other parameters that were varied in the tests included surface (concrete or field), direction (forward or reverse), speed (acceleration or deceleration), motor setting (field or road) and head position (up or down). In addition to varying all these parameters while the windrower was moving, we also studied how the machine behaved when it was stationary (with the hubs locked out). This was done to capture the ideal behaviour of the machine, because we did not need to take into account the slip between the wheels and the surface, and the inertia of the machine. Over 85 tests were performed and each test had three repetitions.

After all the excitement of accelerating and decelerating a vehicle as big as the windrower under the scorching sun in July, it was a welcome change to sit in the freezing Spencer Laboratory to put the collected data into good use. This data was used to fit the wheel speed reaction curve and to develop the machine's transfer function. Subsequently, a proportional-integral-derivative controller was designed and implemented to improve the responsiveness of the driving and steering system as well as to enhance its robustness. The simulated closed-loop response was highly satisfactory, but this did not turn out to be true on the actual machine due to unmodelled second order dynamics. Hence, further on-site tuning were performed in order to achieve a satisfactory level of closed loop step response.

During the first six weeks, we stayed in a hall of residence together with the American students who, like us, were doing research there over the summer. To ensure that the students have a good balance between work and play, the undergraduate research office at University of Delaware had assigned two local students to arrange visits every weekend to nearby cities, which included Philadelphia, Baltimore, Washington DC and New York City.

For the last two weeks, home-stays were arranged for all the IC students. This gave us unique opportunities to gain some insight into American family life, which is quite different to what Hollywood portrays to the rest of the world. It truly surprised me to see some American families holding on more strongly to traditional family values than many Asian families I know of.

The beauty of this exchange programme lies in the combination of hands-on engineering and cultural experiences. The eight weeks spent in Delaware will always remain part of my most cherished memories.

Yuen Yoong Leong won the Old Centralians Trust Jessel Rosen Research Award, awarded to a first or second year Electrical Engineering student to pursue a research topic of not less than six weeks in the summer vacation. In 1999, she won the YMS Short Papers evening competition run by the IEE.

Back To Top

This site uses cookies to track visitors. To stop this happening change the settings on your browser to reject cookies.