DARPA urban challenge

Jean-Charles Bazin

CC500-GroupA - Who has never dreamt of having a car that could drive fully automatically? Making this dream come true is actually the goal of the amazing DARPA urban challenge. The international participants are asked for building an intelligent robot that can drive autonomously in a completely unknown environment by using cutting-edges technologies, such as robot vision, laser, GPS, artificial intelligence, etc... No later than at the end of this week, they have to submit a qualifying video proposal that clearly presents the navigating abilities of their robot. This crucial step is a fabulous opportunity to introduce the DARPA urban challenge in this hot news.

This presentation is divided into three main parts. First, I will explain what the DARPA challenge is and introduce the goal and results of the previous editions of DARPA competitions. Then the brand new ’07 urban challenge will be presented, and finally, the current status of one team (KAIST-Upenn) will be analyzed to show some key technologies involved in this worldwide competition.

First of all, the acronym DARPA stands for US Defense Advanced Research Projects Agency whose aim is to develop future technology for military applications. In order to accelerate research and strengthen relationships with universities, DARPA has created the DARPA challenge. The first edition of DARPA challenge took place in 2004. The cars should travel more than 220 km in the Mojave Desert (USA) without any human interventions. Due to the unbelievable difficulty of such a task, no team managed to complete the distance, and worst, the best team (CMU) has traveled only 12 km. After one year of intensive work, no less than four teams have succeeded in driving the 200 km which was far from being expected. The best car, by Stanford, has completed the distance in only seven hours and won the $1 million first prize.

A brand new edition of DARPA challenge is scheduled for November 2007. Whereas ’04 and ’05 competitions took place in the desert, ’07 edition will be held in urban environment. Therefore, the difficulty is still much higher. Indeed, the cars have to not only take other moving vehicles into account but also obey all traffic regulations. For example, if a car detects an obstacle, it has to modify its trajectory to avoid the collision. A very complex situation occurs when a car arrives at a crossroad and has to check whether no other car is coming before crossing. In regards to the complexity of this edition, the first prize has been increased to $2 million dollars.

Framework of the technologies involved in the DARPA urban challenge: robust GPS communication, robust digital map localization, car and moving object detection

In order to introduce some key technologies involved in the urban challenge, this part focuses on the current status of the Ben Franklin team. This team is the result of an intensive collaboration between GRASP lab at the university of Pennsylvania and RCV lab at KAIST. The project has been divided into two parts. The Korean group is focused on the sensors to gather information from the environment. For example, the goal is to automatically detect road lane markings, moving cars, obstacles and use GPS data, and finally match the results with digital map. These tasks are mainly based on computer vision and artificial intelligence and are very complex. Indeed, for instance, how a machine can detect marking lanes automatically with a 100% accuracy even in presence of shadows, rainy or sunny weather or when lanes are of different colors. Moreover some roads will not have lane marking so the car also has to detect this situation. The US group is working on the car dynamic, that is to say how to control the car, drive it faster or slower, change the direction, etc... It is mainly based on information provided by the Korean team and a small navigation error can lead to a car crash and have dramatic consequences for the expensive embedded equipments. .

To conclude, DARPA challenges have encouraged researchers all over the world to develop new technologies and solve robotic problems that seemed to be impossible even three or four years ago. I am convinced that ’07 edition will bring amazing systems that will set back robotic limits still further.

References:

- http://en.wikipedia.org/wiki/DARPA

- http://en.wikipedia.org/wiki/Urban_challenge

- http://www.benfranklinracingteam.org/

Hot news test example

Can't Knock It Down

Julie J. Rehmeyer

CC500-groupA- The "Comeback Kid" is a wooden toy with an intriguing property: No matter which way you set it down—on its head, for example, or on its side—it turns itself upright. Two factors account for this: the object's shape, and the fact that the bottom of the toy is heavier than the top.

Set the Comeback Kid in any position, and it will turn itself

upright. Theoretically, it's possible to balance the figure on its

head, but the slightest breeze would knock it over and restore

it to its upright stance.

Give mathematicians such a toy, and they're liable to turn it into a math problem.

Mathematicians Gábor Domokos of the Budapest Institute of Technology and Economics and Péter Várkonyi of Princeton University wondered if they could make an improved version that wouldn't require the weight at the bottom to right itself. Could the shape of the object alone be enough to pull it upright?

They started experimenting with flat toys cut from a piece of plywood. They cut out shape after shape and found that the edges of each shape had at least two stable balance points. In addition, each shape's edges had at least two more points on which the mathematicians could balance it if they were very, very careful, but the slightest breeze would knock it over. They refer to those as "unstable balance points." (Similarly, it is possible, barely, to balance the Comeback Kid vertically on its head.)

Eventually, Domokos and Várkonyi managed to prove mathematically that for any flat shape, there are at least two stable balance points and at least two unstable balance points.
Next, the pair began to investigate whether all three-dimensional shapes have at least two stable and two unstable balance points. They tried to generalize their two-dimensional proof to higher dimensions, but it didn't hold up. Therefore, it seemed possible that a self-righting three-dimensional object could exist. Such a shape would have only one stable and one unstable balance point.

They looked for objects in nature that might have such a property. While Domokos was on his honeymoon in Greece, he tested 2,000 pebbles to see if he could find one that would right itself, but none did. "Why he is still married, that is another thing," Várkonyi says. "You need a special woman for this."

Eventually, the team managed to construct an object mathematically that has just one stable and one unstable balance point. The figure is like a pinched sphere, with a high, steep back and a flattish bottom. They sent their equations to a fabricator, who constructed the object. Várkonyi now keeps it in his office. "People like playing with it," he says.

Domokos and Várkonyi used mathematics to design this self-righting object.

Once the pair had built their self-righting object, they noticed that it looked very much like a turtle. They figured that wasn't an accident, since it would be useful for a turtle never to get stuck on its back.

The shape of the Indian Star Tortoise is similar to the self-righting

object that Domokos and Várkonyi created. When turned onto its

back, its shape helps it come close to flipping over without effort,

but the turtle needs to give itself a little boost by kicking its legs.

Now, Domokos and Várkonyi are measuring turtles to see if any of them are truly self-righting, or whether the turtles need to kick their legs a bit to flip themselves back upright. So far, they've tested 30 turtles and found quite a few that are nearly self-righting. Várkonyi admits that most biology experiments study many more animals than that but, he says, "it's much work, measuring turtles."

The mathematicians still face an unanswered question. The self-righting objects they've found have been smooth and curvy. They wonder if it's possible to create a self-righting polyhedral object, which would have flat sides. They think it is probably possible, but they haven't yet managed to find such an object. So, they are offering a prize to the first person to find one: $10,000, divided by the number of sides of the polyhedron.

It sounds like a tempting challenge, but there's a catch: Domokos and Várkonyi are guessing that a self-righting polyhedron would have many thousands of sides. So the prize might only amount to a few pennies.
References:

Domokos, G. 2006. My lunch with Arnold. Mathematical Intelligencer 28 (Fall):31-33. Reprint available at http://www.szt.bme.hu/Munkatrs/domokos/cikk_archiv/99/final/99.pdf.

What is a CPU?

What is a CPU in computer science?

JC Bazin -A processing unit (CPU), or sometimes simply processor, is the component in a digital computer that interprets computer program instructions and processes data. CPUs provide the fundamental digital computer trait of programmability, and are one of the necessary components found in computers of any era, along with primary storage and input/output facilities. A CPU that is manufactured as a single integrated circuit is usually known as a microprocessor. Beginning in the mid-1970s, microprocessors of ever-increasing complexity and power gradually supplanted other designs, and today the term "CPU" is usually applied to some type of microprocessor. The phrase "central processing unit" is a description of a certain class of logic machines that can execute computer programs. This broad definition can easily be applied to many early computers that existed long before the term "CPU" ever came into widespread usage. However, the term itself and its initialism have been in use in the computer industry at least since the early 1960s (Weik 1961). The form, design and implementation of CPUs have changed dramatically since the earliest examples, but their fundamental operation has remained much the same. Early CPUs were custom-designed as a part of a larger, usually one-of-a-kind, computer. However, this costly method of designing custom CPUs for a particular application has largely given way to the development of mass-produced processors that are suited for one or many purposes. This standardization trend generally began in the era of discrete transistormainframes and minicomputers and has rapidly accelerated with the popularization of the integrated circuit (IC).

example of CPU

The IC has allowed increasingly complex CPUs to be designed and manufactured in very small spaces (on the order of millimeters). Both the miniaturization and standardization of CPUs have increased the presence of these digital devices in modern life far beyond the limited application of dedicated computing machines. Modern microprocessors appear in everything from automobiles to cell phones to children's toys.