California Trip 1997

Bill works in the vision lab of the Vision and Earth Science Applications group at JPL. His work area is situated on the other side of an eight-foot wall of work stations and video monitors where the first pictures were received from the Sojourner Mars mission this past summer.

Bill’s work involves the production of next-generation rendering protocols for robotic planetary rovers (small, six-wheeled robotic exploration vehicles). The idea is to automate some of the navigational decision-making capabilities of the rover. The underlying system developed by JPL for this purpose is known as "S.O.A.P.", which is short for Sense of Active Presence. Using SOAP data (mathematical and graphical representations of the terrain in the immediate vicinity of the rover) rover drivers have been able to determine clear-path and target-object information.

The rendering application, previously written in raw Open-GL, is gradually being converted to an SGI-based program with Motif interface. Bill needed to learn Performer API to upgrade the program so that the rover drivers could move it real time within its environment—a work in progress. Bill had to learn Performer API and C++ on the job.

He said that it is common for people from his group to go to work for the Disney Studios after a few years at JPL.

Bill’s supervisor, John Wright, has little daily interaction with Bill, however, this doesn’t seem to concern Bill very much. For his part, Wright seems very pleased with Bill’s work. Bill described his living conditions as "great". He lives in the guest house of a JPL employee.

Marc works in an adjoining laboratory to Bill Feth’s work area. His group is the Digital Image Animation Laboratory group, made up of about a half dozen students and scientists. This group is responsible for creating things like the 360° HDTV images of the landscape surrounding the Sojourner Mars lander. Here is a picture of that lab.

Marc’s primary project is to work with an application called Vista Pro which he uses to manipulate satellite images and other data taken of near-Earth atmospheric conditions. Using his knowledge of JAVA, he has created several interactive presentations of the data. For example, the space shuttle collects CO₂emission data from a ground based box called the Alaska Synthetic Aperture Radar Facility. Marc has been given the job of creating a web-based interface which will allow scientists and students to explore the data through a graphical rendering. Think of a 3-D weather map that shows layers of clouds and other atmospheric information. Then imagine that this graphical representation is provided with an interface that allows users to click on sub-areas and zero in on more specific images and or data. One set of images might show temperature layers, while another shows the same space according to its CO₂levels.

A major goal of this work is to enhance the lab’s ability to make this kind of information available to educators and students worldwide.

Marc’s supervisor is Charles Thomas, a geographer. I enjoyed talking with Charles, not only about Marc and his work (which Charles was both pleased and impressed with) but also about how he, Charles, came to JPL and how he viewed his own work at the lab. He talked about the close-knit atmosphere of those working at the lab and about the deep sense of purpose and exploration that he and others feel that their work provides them with.

This warm, collegial atmosphere seemed to trickle down to the students as well, who seemed generally pleased with their work and living arrangements. Marc lives with two other Cornell Coop students (Bawb Sanii and Brian Camilli) in an apartment nearby. They ride a shuttle bus to work each morning.

Peimin works with the Digital Signal Research Group. This group looks at problems which involve earth-based, command and control systems of un-manned spacecraft. Transmission and reception of digital signals are routed through special optical modems built by Motorola. Peimin’s work involves the testing of RS encoded source signals on the receiver end of the system (the earth-based controllers). Peimin was able to visit the actual ground station where the controller is based. This station was some distance from Pasadena, located (I think) in the mountains to the east.

Small bursts of information are passed back and forth between such earth-based control stations and distant flight platforms. The digital signals being tested by Peimin are carried on at a data transfer rate of only 100 kilobit/second. In a report to his Dr. Yan Peimin wrote:

Despite the obvious limitations of the computer, real time encoding was still achieved.

Previously there had been no SCSI interface for the system. Peimin also worked on creating this interface.

Peimin’s assigned supervisor, Victor Vilnrotter served as more of an acting, or official, supervisor of record. His real supervisor was instead, Tsun-Yee Yan, who happened to be out of the office on the day I visited. Peimin was happy with his living arrangements and social life in Pasadena, however, he had two concerns to pass along to the Coop Office. The first was that access to Cornell’s Pre-enrollment system was difficult and frustrating (a sentiment echoed by other students). And second, he was finding it difficult to look for spring housing in Ithaca from such a distance and wondered if the Coop Office could be of more help in this regard.

Victor is working on planning for Deep Space 4, one of several "Deep Space" missions scheduled to be launched in the first years of next decade. Deep Space 4 will encounter a near earth asteroid. Once the spacecraft attains a stable orbit around the asteroid it will begin to map its surface. This will take as much as 151 days. With surface data acquired, the mission then will focus on finding optimal landing sites. A site will be selected and landing will occur. The spacecraft will need to be prepared for a wide range of possible surface conditions. A successful landing will allow the lander to probe the asteroid’s surface in order to take samples.

The final phase of the mission will involve the launch of return craft from the lander. The return ship will carry samples from the asteroid back to earth.

Victor’s role in the mission is to help detail the engineering master plans for an entire stage of the deep space mission. Complex flow charts and planning algorithms are used to map out every detail of the mission. A major stage of the mission is the cruise stage during which the spacecraft flies to/from an asteroid. This was described to Victor as "creating a behavioral model of the mission’s cruise stage". The software application used by JPL for this purpose is a planning program called Foresight. Victor’s responsibility entailed coordinating the integration and communication between the Foresight model and other related systems and models. In order to work with this software Victor has to use a workstation located in a lab one floor below his office area. Although he described having to time share on the lab machines as "frustrating" he also said that he understood that his group wasn’t in a position to just buy him his own $60,000 workstation.

Victor’s supervisor, Grace Tan Wong was not available for me to meet but he said that they had a good working relationship. He said that he doesn’t work with her directly very much. He felt his placement was well planned and appreciated all of the trouble people had gone to help him get up to speed.

Victor bemoaned the lack of good public transportation in the area (without a car you’re more or less stranded). He likes where he lives though and has fun taking advantage of local shopping areas and entertainment.

Brian’s work involves testing of a Gallium Arsinide transistor being used in a programmed microelectronic stepper motor. These tests measure things like durability under certain environmental conditions like high/low temps. The intent is to isolate non-destructive mechanical systems. These MEMS, short for Micro Electro Mechanical Systems, must be thoroughly tested prior to fabrication.

The head of this small group is Dr. Qui Kim. Brian’s work, however, is being supervised by Sammy Kayali. Kaylli was busy and unable to speak with me at length, however, he was supportive of Brian’s work and seemed pleased that Brian was working with the group.

Brian said that two Cornell courses were particularly helpful to him in his work at JPL. EE 315, dealing with quantum mechanics, and EE 493, which is a MEMS course.

Brian talked fondly of his trip out to California. He drove with two other students at JPL’s expense and got to see the country on the way.

In a shared office facility on the other side of Pasadena, this JPL group works on automation of various support systems for spacecraft. Deep space control stations must be placed around the earth at 120° apart from each other, covering a full orbital sweep around the globe. The group located in this office is responsible for a $25 million upgrade of the existing platforms. The system is the "front-end" of the antennae and related subsystems (like telemetry, transmitter, receiver, and ground tracking sites). NASA tracking sites are large radar dishes located in places like Goldstone, CA.

Brian’s work involves turning existing protocols into automated operational functions. Previously this had been done in something called Automated Motion Language Operation. Brian’s task is to use TCL (Tool Command Language) to automate various protocols, like point-to-point predictions. Another task was to create an editor interface for system so that others could make changes in the program as needed.

While Brian didn’t feel that this work was really "in his field", he said that he enjoyed his co-workers and found the overall environment to be much better than that of industry. He said that he would strongly consider working for JPL following graduation. He also is thinking about business school.

I met with Ava Bokor (a mathematician) rather than Brian’s official supervisor, Hu Henry. Bokor works more closely with Brian and Henry did not seem to be available. She said that Brian was "so smart that it is frightening". She was amazed that he was able to pick up Unix, TCL/TK, and x-Windows so quickly. She admitted that the conversion called for the experience of a seasoned veteran, but that to her great surprise Brian was able to do a super job with little or no formal background in the system or related applications.

Babak’s group is involved with earth-to-spacecraft and spacecraft-to-earth optical communications. Highly sensitive positioning information must be exchanged between earth-based stations and distant spacecraft in order for them to acquire each other in the vastness of space.

Babak was presented with some very specific constraints that current systems experience due to the nature and motion of space flight. For example, JPL spacecraft employ a gimbal mechanism to offset the major positioning variances caused by the irregular motions along a flight path. A gimbal is defined as a "suspension device whose orientation can be changed without affecting the attitude of the body being suspended." Once these macro variances are adjusted for, the system uses very sensitive piezoelectric detectors to lock onto the sought after signatures of laser light coming from earth. Babak’s job description was to construct a workable test platform for these ultra-sensitive devices and then solicit potential interest from manufacturers of related devices.

By the time I had arrived in early November, Babak had already succeeded in building a viable test environment. His tests led him to conclude that he would need a test device capable of very, very fine sinusoidal outputs. The device in question would have to have translation capacity in the 2600 baud range carrying approximately 8 gigabits of streaming information. Unlike the other coop students at JPL, Babak’s work was very hands on with lots of shop time required. He seemed to enjoy this aspect of his work.

Babak’s supervisor, Muthu Jeganathan, was a somewhat quiet fellow who had very little to say about Babak’s work. He seemed almost indifferent to the subject. Babak’s own notes however suggest that Muthu was very supportive and frequently checked in on him.

It was obvious to me that Babak was enjoying the LA area, especially given the fact that he is involved in acting. He describes his living quarters as "a fine place that’s a little far from work, but manageable." Babak lives with Marc Casalaina and Brian Camilli.

Kelly has an office in the same building that Babak does. But unlike Babak, Kelly spends most of her time in an office setting. She has been given the task of creating an interface program for calculating earth-based locations for ground satellite communicators.

One of her first tasks was to read a number of papers on the kinds of research being done by her group. In particular she needed to become acquainted with the Spacecraft Orbital Analysis Program (SOAP) which was developed by a company called the Aerospace Corporation.

Kelly wrote a computer program that uses SOAP data to create a cloud-cover map of the globe. The objective is to allow scientists who are working on the new, multi-nation spacestation project to quickly visualize the placement of potential ground telescopes around the planet. The configurations typically involve sets of three tracking ground stations located in such a way that an orbital object, like the space station, will come under constant observation as it circles the globe. The sections of space that need to be covered by Kelly’s program are shaped like cones with large domes on top. The angle from zero degrees (level) on earth is a steep 20°, which is quite a bit smaller than that of the shuttle tracking stations which need be only 4° to capture the low orbit of the shuttle.

Kelly’s supervisor, John Sandusky, took a few moments to say hello to me. He was pleased with Kelly’s work and supportive of the whole coop initiative.

Kelly writes about her living situation: "We have a six month lease and are looking to rent it for the spring semester." Kelly describes her living situation as "good" and lives with another coop student nearby JPL.

Annika’s main task is building tie-ins for a network analyzer. This particular network analyzer is used for printer applications at HP. The actual printer application code is written in C++. The specific code that Annika is working on is used for system measurements and communication protocols. For example, finding default printers and making adjustments. The network analyzer becomes a filter of sorts, transmitting or reflecting specific protocols. A big part of Annika's work is to write help files to go along with the analyzer. The overall project that Annika's group is working on is for wireless communications giant Ericsson.

Another project that Annika is involved with is called Component Object Modelling. Using a Microsoft application, Annika is creating an interface for a project printer. Test equipment is developed to be used for other R&D needs at HP. Eventually this equipment becomes standardized and is incorporated on a production basis. Overall, the object of the work is to create reliable solutions to systems problems.

Annika's direct supervisor is Allen Kashawa, who seemed very pleasant and happy with Annika's work. Annika mentioned that she was impressed with the whole work environment at HP and found the culture there to be especially focused on time management issues.

Took "Red Eye" to LA. Liked Pasadena a lot. Wonderful atmosphere and good food. Stayed at the DoubleTree Inn. Nice gym in the hotel.

Courtyard Hotel in Marina Del Ray was over 2 miles from Venice beach. Much farther than I thought it was.

Had an alumni reunion dinner in Palo Alto. Dined at Nola's, which was a lot of fun.

Nolas2.JPG (18299 bytes)

Nola's in Palo Alto, California. Pictured from left to right: Kelly Chai '90
(Pixel Translations), Arjun Kapur '92 (Intel), Dan Jenkins '82 (Cornell),
kneeling, Nisheeth Ranjan '96 (Netscape), Scott Benson '90 (Neuro
Studios), Bill Garrett '93 (Silicon Graphics), Reza Beforooz '97 (Stanford),
Professor Mike Godfrey (Cornell).

Not pictured, but attending that evening were: Lisa Camesano '92 (Sun), John
Byers '91(Berkeley), Aravind Narasimhan '96 (Sun), Hersh Reddy '96 (Flat
Line), and Jose Machuca '96.

Friday morning I drove up to Santa Rosa (about an hours drive). Stayed at a hotel called The Flamingo. Nice health club nearby but the hotel was kind of weird.

On Saturday I tried making one of the boat trips out to Alcatraz, but everything was booked (I had forgotten about this).

Last updated November 23, 1998. For more information email me at jenkins@cs.cornell.edu .

Brian Camilli ’99 (CS)	Bill Feth ’99 (CS)	Kelly Shea ’99 (EE)
Marc Casalaina ’99 (CS)	Annika Karlsson ’99 (CS)	Brian Stark ’99 (EE/CS)
Peimin Chi ’99 (EE)	Babak Sanii ’99 (AEP)	Victor Tao ’99 (EE, switching to CS)