Far from ordinary ordnance removal

. Their solution mimics many of the systems in manufacturing High explosives, students and off-road excavators could form a potentially lethal combination – the ultimate rag week high perhaps. However, put a real-time control system, remote operation and the backing of one of the world’s largest automation companies behind them and a deadly serious problem is close to being remedied. The Virgina Tech students were presented with the dilemma of how to remotely dig holes in the ground at the US Naval Surface Warfare Centre in Dahlgren, Virginia; the reason being that a large amount of dangerous, high-explosive ordinance had been buried in the ground after it had reached the end of its ‘useful’ life. After numerous ideas were discussed and rejected the resulting solution was a remote controlled vehicle of enormous proportions – for some reason Thunderbirds springs to mind. Not the usual fodder for Eureka you may think, but when you consider the hydraulics and electronics normally found on an off-highway vehicle could quite as easily find them selves in a manufacturing environment doing a completely different job then it all comes into focus. Also, given the fact that the remote controlled behemoth was plying its trade in a field in Virginia while being driven from a virtual cab in Austin Texas, then data acquisition, two-way communication, visual signals and 1,200 miles worth of remote operation also come into play. Not your usual, run-of-the-mill automation application but a powerful demonstration of what can be achieved with the right software, the right hardware, supportive companies and bit of imagination. The vehicle in question, a heavily modified Case Construction Equipment CX160, is required to safely uncover, handle and remove ordnance which in some cases, according to the US Navy, could include bombs weighing as much as 500 pounds. All of this is to be achieved from the safety of a shielded trailer some 1,000 yards from the excavation site. The demonstration witnessed by Eureka was a little extreme, with the excavator and controller being 1,200 miles apart, but it did ably demonstrate the capabilities of web-based control. The student team also won first place in National Instrument's NI Week 2002 'Best Application of Measurement and Automation' contest. A number of features were immediately noticeable on the outside of the CX160 namely the antenna array, the control box, a 'thumb' and the cameras – the 'thumb' being added to offer an opposable grip to the bucket. On the control front, the Virginia Tech students modified LabVIEW software from National Instruments to enable them to control remotely the excavator via radio modems and the Internet. The human controller maintains visibility of the area being traversed and dug through the use of three cameras. "These cameras give us the ability to see all angles of the excavator," explains mechanical engineering graduate Chris Terwelp. "One camera has a pan-and-tilt module; and we will be adding a zoom feature as well." Figuring out how to safely operate a 35,000 pound, 110-horsepower excavator by remote control was a major focus of the project. "The issue was how to make sure this thing didn't get away from us," said mechanical engineering professor Al Wicks. To this end the students created an emergency shutdown system using a line-of-sight radio link to the excavator. If a continuous tone being broadcast to a receiver on the excavator is interrupted, the machine automatically shuts down. Basic excavator operation is achieved by reading analogue voltages from control input devices at the remote station. Industrial joysticks and foot pedals provide control for the excavator's six degrees of freedom, swing, boom, arm and bucket as well as left and right track, and four buttons on the joysticks facilitate the pan and tilt capabilities of the camera. An National Instrument's Field Point analogue voltage input device reads control data (at rates up to 350Hz – speed is crucial for reliable digital control) and a real-time series network module initiates the data acquisition process. The analogue voltage is then sent, as a string, across the network for implementation on excavator-side Field Point modules. A discrete input module controls ignition and also allows the operator to remotely toggle the power of the vision system and the work lights. To ensure the optimum running of the vehicle, the operator can monitor all parameters such as fuel level, water and oil temperature and warning messages in real time via LCD screens next to the joystick. Design engineers at Case, in conjunction with engineers from Sumitomo, developed software to report this data via an RS-232 port located on-board the excavator’s computer. An unused Field Point Ethernet connection is used to monitor the status of the excavator with the data being processed and sent back across to the remote site where the proper action can be taken. If a critical error occurs, vehicle control is terminated and warning messages are displayed to the user. A Windows based PC running a third LabVIEW virtual instrument (VI) monitors the excavator by reading information published on the remaining remote-side ethernet port. The PC is used mainly as a monitoring device, reducing 'strain' on the real-time processor. It can also have limited control of the vehicle. The excavator can be monitored through any PC with either LabVIEW or a web browser. "This is a great example of good work done on time," said Wicks. "These students really showcased their skill sets and were able to work together to retrofit the excavator." Wicks and mechanical engineering professor, Charlie Reinholtz, supervised Terwelp and the five undergraduate students who worked on the project. In putting together a team for the project, Wicks and Reinholtz looked for students with an interest in electronics and computer interfacing, as well as a knowledge of LabVIEW. Completing the project in only six months, the team constructed a functioning system, met deadlines, and stayed within their budget. Pointers Remote control is normally over 1,000 yards but the project was demonstrated over 1,200 miles Full visibility is provided by three cameras with control from a dual-joystick control board Real-time operation is made possible using advanced control infrastructure