The National Institute of Standards and Technology (NIST) is seeking university teams to participate in a May 2010 Virtual Manufacturing Automation Competition to simulate an industrial robot performing a common but complex shop floor task—stacking odd-lot boxes on a shipping pallet. NIST, the engineering society IEEE and Georgia Tech are co-sponsoring the competition. The winning team will run part of their simulated task on a robot at the International Conference on Robotics and Automation 2010 in Anchorage, Alaska in May as part of a demonstration of seamless transition from simulation to real-world systems.
"To participate, the teams need a computer gaming engine that is available for about $10," says engineer Steve Balakirsky, adding that "from there they can use existing computer code, or create their own, to develop a simulation of a robot picking up boxes of various sizes and weights from a conveyor belt and arranging them on a pallet for shipping." NIST's interest, Balakirsky says, is in devising "performance metrics" that can be used to determine what makes a "good" mixed pallet.
The "mixed palletizing" task is a current manufacturing research interest because plants that produce multiple products often have to ship a variety of items to a single location. The problem is familiar to anyone who has packed a variety of gifts into a larger box for shipping. A sound solution could lead to more efficient delivery of mail, food and other wholesale items.
Building a mixed pallet is an efficient method of transporting the goods, but programming a robot for this task requires knowledge of robotics, mobility, mapping and scheduling in a manufacturing environment. Additional consideration in creating the mixed pallet are package density, pallet stability and how to know which packages need special care due to weight or fragility. The current state-of-the-art use of robots to place products on pallets involves stacking a predetermined number of boxes of the same size and weight. The challenge posed by the VMAC competition will stimulate advances in many areas of robotic algorithms, ranging from the perception of the boxes' geometry, the grasping and positioning algorithms, and the overall planning procedures for intelligently configuring the mixed pallets based on the available boxes.
This simulation-based competition allows students to engage in real-word research that is ideal for learning robotic architectures, motion planning systems and multi-robot control, Balakirsky said.
The deadline for entering the competition is Feb. 15 and the competition is May 2-3. More information about the competition, including a simple example of a robot simulation, can be found at www.vma-competition.com. Contact NIST researchers at robosim [at] nist.gov with questions.