You know it's time to update when. . .
I figured it was about time to make an update, considering how long it's been since my last one.
I am tired. Very, very tired. I-can't-wait-for-winter-break-and-it's-over-a-month-away tired, but at least I'm getting a lot accomplished in terms of the autonomous robot project, and even more so at work.
Autonomous robot project:
Even though I am a mechanical engineer, I am on the software team for this robot. This means I spend about a third of my time helping out the mechanical team, the electrical team, or both and the remainder of my time developing, debugging, and refining the software that runs the bot.
My helping of the mechanical group has consisted of design and analysis help, pointing out flaws in the design that they should correct, helping with component and material selection, planning for manufacture of custom components, etc. Helping out both, I wrote an Excel program that tells us what electrical current requirements we will run into with the motors, in terms of peak, continuous, and average current. In a related note, I helped the mechanical team in the preliminary power calculations for narrowing down the list of feasible motors to ones which would actually work. Helping just the electrical team, I explained how the emergency stop switch should be hooked up and how to get around the fact that it nominally couldn't handle as much current as we want to put through it. Namely, that they should get a 4-terminal switch and split the current between 2 pairs of terminals instead of running it all across a single pair. It seemed obvious to me, but they couldn't figure out how to get around it at first.
When I joined the project, the basic motion control algorithm had been developed by another member of the team. More specifically, it had been lifted almost in its entirety out of the book we are using for reference. Since then, I have helped enter the algorithm into Labview and have added multiple features to it, including an algorithm that allows us to control the top speed of the bot and makes it slow down through the turns and adding path-following ability to it. All of these I developed and implemented completely on my own.
Technically, the motion control is not on my list of responsibilities for the software, but I saw the need and filled it. My list of responsibilities is merely the following: map building (including obstacle detection and plotting), localization (including Kalman filter design), cognition, decision making, and path-finding. I have picked out the schemes to use for each of these. How I implement them will depend on what sensors we end up being able to buy, which depends on how much money and/or equipment we get from sponsors.
Speaking of sponsors, we have a preliminary design review presentation before multiple potential sponsors on Thursday of this coming week. I figured we should have something to "wow" the sponsors and make them say, "These guys have a real chance of winning if they can do this much with almost no money! Let's dig deep and give them a huge budget!" To that end, I had someone else on the team videotape the bot first going in a circle, then making a figure-eight all on its own. I figure they will see this and be somewhat impressed, but not wowed. Therefore, I will let them digest the fact that we just proved that we can make the bot follow a path once the path-planning software gives it to the motion control program, then show them the next video. In this video, I have made the bot actually spell out IGVC (for Intelligent Ground Vehicle Competition, the contest in which we are going to take part). If anything from our tech demo is going to impress them, I think this will.
Work:
I have been spending quite a lot of time at work, including overtime, to get my work done on the programs on which I have been working. The first of these is through our offices in Canton, Massachusetts. My job on this program has been to design a new fiber-optic gyroscope (for those of you who don't know, it's a sensitive rotation sensor). It is the first one I have ever designed, and had to be thinner than any of the 400-meter fiber optic gyroscopes we already had. One of the requirements was that it be able to use a particular coil known as the "Atlas" coil as well as being compatible with two other coil designs. I completed my design a couple of weeks ago and the optical engineer from Canton ran an analysis on it to determine its performance level. According to his analysis, it will be, by a significant margin, the highest-performing "Atlas" gyro ever built. This is in comparison to the other 6 which have all been developed by the PhD's in Canton. I was, and still am, very proud of myself for this. Especially so, given that we had only 8 weeks to develop the entire system (compared to 6 months average) and I finished almost the entirety of my gyro design in only three weeks, after multiple changes in requirements (I can think of 5 right of the top of my head) from the customer.
As a side note, I was working on the program with an experienced mechanical designer. His responsibility was the electronics box, while mine was the gyro. He retired a couple of weeks ago, and there was a retirement luncheon for him and several other guys from our area who were all retiring at about the same time. Until he retired, he had a reputation as being the best mechanical designer at our facility (out of about 30 or so) and for being extremely capable with the SolidWorks solid modeling program. For reference, when I designed my wind-up powered gyro calibration mechanism last year, the boss I had at the time made the comment that this designer was probably the only other person at the facility who could have done something like that, with one or two others having some chance. Well, on the afternoon of the day of the luncheon, I was invited to the department director's weekly staff meeting (for reference, he's my boss' boss' boss). At the meeting, he had everyone in the room talk about what they are working on and how it's going. I mentioned my progress on my gyro design, and one of the managers on par with my boss piped in, "I was talking to [the designer] about you at the luncheon today, and he told me, 'That kid is a SolidWorks genius!'" I just sort of blinked, stunned, and felt my face turning bright red, "Wow. That is really nice to hear. Thank you for telling me." I thought, but did not say, 'And thank you even more for saying it so blatantly in front of my boss, his boss, and his boss.'
The second program has just started, and I was assigned to it in large part due to my performance on the above-mentioned program. In this program, my job is to first develop the design of the initial test-bed for a new type of gyro, a nuclear magnetic resonance gyro (NMRG), and then miniaturize it. I am one of only 4 people at our facility working on the design, and I am the sole mechanical engineer and doing the entirety of the mechanical design myself. The others on the team are a physicist, an optical engineer, and an electrical engineer.
This would be the continuation of a program that started in the late 60's and ended in 1982, when they had the test gyro down to approximately a 12-inch long by 4-inch diameter cylinder. I say that it would be the continuation of this program except for the fact that all components except for 4 core cells, all drawings, and almost all design and scientific documents about the previous iteration of the program were literally thrown out some time during the last 23 years. Therefore, we are starting nearly from scratch. All we have to go on are a few pictures, a couple of hand-written notebooks with incomplete data and references to no-longer-existing documentation, and the memories of a physicist who is about to retire and who worked on the program in its previous incarnation.
Based on what I know of the requirements, I have the bulk of my conceptual design for the first test bed done, minus selecting and making changes to accommodate for a few commercial-off-the-shelf components. So far, it's about half the size of a Coke can, and should grow to about 3/4 the size of a Coke can by the time the extra components are added. It can accommodate not only the left-over core cells but also new cell designs down to the expected 1mm size without difficulty. We are outsourcing the cell designs, and have a collaboration of different organizations working on them, including UCLA, CalTech, and a company that makes the cells for atomic clocks. After I select these components, I am going to have to develop the coil apparatus that will give us as close as possible to a completely uniform magnetic field throughout the core cell.
Once we prove that we can make a gyro with this technology, it will then become my job to miniaturize it down to 1 cubic centimeter or less. In short, it looks to be difficult as hell and I can say with no sarcasm whatsoever that I am very much looking forward to that stage of the program.
Unfortunately, between work and school, I seem to come home practically brain-dead almost every night. Imagine taking final exams for 8-12 hours, 5-6 days a week for 6 weeks straight, and you will get an idea of how mentally tired I am, and it is starting to really bleed over into physical tiredness. It was already bleeding over the last time Ash_grey_sky saw me, and it has definitely worsened in the last couple of weeks. Thanks to getting ready for the design review on the autonomous robot project, I have had very little free time since Tuesday evening. I had some free time last night, which I spent vegetating in front of the TV for the last half hour of Star Wars Episode III, then playing Guild Wars with Grampablohm and making clear in my conversation with him (through constantly losing my train of thought) just how brain-dead I was feeling.
I think I just proved how tired I am by rambling on about being tired instead of going to sleep and reducing it. On that note, good night.
I am tired. Very, very tired. I-can't-wait-for-winter-break-and-it's-over-a-month-away tired, but at least I'm getting a lot accomplished in terms of the autonomous robot project, and even more so at work.
Autonomous robot project:
Even though I am a mechanical engineer, I am on the software team for this robot. This means I spend about a third of my time helping out the mechanical team, the electrical team, or both and the remainder of my time developing, debugging, and refining the software that runs the bot.
My helping of the mechanical group has consisted of design and analysis help, pointing out flaws in the design that they should correct, helping with component and material selection, planning for manufacture of custom components, etc. Helping out both, I wrote an Excel program that tells us what electrical current requirements we will run into with the motors, in terms of peak, continuous, and average current. In a related note, I helped the mechanical team in the preliminary power calculations for narrowing down the list of feasible motors to ones which would actually work. Helping just the electrical team, I explained how the emergency stop switch should be hooked up and how to get around the fact that it nominally couldn't handle as much current as we want to put through it. Namely, that they should get a 4-terminal switch and split the current between 2 pairs of terminals instead of running it all across a single pair. It seemed obvious to me, but they couldn't figure out how to get around it at first.
When I joined the project, the basic motion control algorithm had been developed by another member of the team. More specifically, it had been lifted almost in its entirety out of the book we are using for reference. Since then, I have helped enter the algorithm into Labview and have added multiple features to it, including an algorithm that allows us to control the top speed of the bot and makes it slow down through the turns and adding path-following ability to it. All of these I developed and implemented completely on my own.
Technically, the motion control is not on my list of responsibilities for the software, but I saw the need and filled it. My list of responsibilities is merely the following: map building (including obstacle detection and plotting), localization (including Kalman filter design), cognition, decision making, and path-finding. I have picked out the schemes to use for each of these. How I implement them will depend on what sensors we end up being able to buy, which depends on how much money and/or equipment we get from sponsors.
Speaking of sponsors, we have a preliminary design review presentation before multiple potential sponsors on Thursday of this coming week. I figured we should have something to "wow" the sponsors and make them say, "These guys have a real chance of winning if they can do this much with almost no money! Let's dig deep and give them a huge budget!" To that end, I had someone else on the team videotape the bot first going in a circle, then making a figure-eight all on its own. I figure they will see this and be somewhat impressed, but not wowed. Therefore, I will let them digest the fact that we just proved that we can make the bot follow a path once the path-planning software gives it to the motion control program, then show them the next video. In this video, I have made the bot actually spell out IGVC (for Intelligent Ground Vehicle Competition, the contest in which we are going to take part). If anything from our tech demo is going to impress them, I think this will.
Work:
I have been spending quite a lot of time at work, including overtime, to get my work done on the programs on which I have been working. The first of these is through our offices in Canton, Massachusetts. My job on this program has been to design a new fiber-optic gyroscope (for those of you who don't know, it's a sensitive rotation sensor). It is the first one I have ever designed, and had to be thinner than any of the 400-meter fiber optic gyroscopes we already had. One of the requirements was that it be able to use a particular coil known as the "Atlas" coil as well as being compatible with two other coil designs. I completed my design a couple of weeks ago and the optical engineer from Canton ran an analysis on it to determine its performance level. According to his analysis, it will be, by a significant margin, the highest-performing "Atlas" gyro ever built. This is in comparison to the other 6 which have all been developed by the PhD's in Canton. I was, and still am, very proud of myself for this. Especially so, given that we had only 8 weeks to develop the entire system (compared to 6 months average) and I finished almost the entirety of my gyro design in only three weeks, after multiple changes in requirements (I can think of 5 right of the top of my head) from the customer.
As a side note, I was working on the program with an experienced mechanical designer. His responsibility was the electronics box, while mine was the gyro. He retired a couple of weeks ago, and there was a retirement luncheon for him and several other guys from our area who were all retiring at about the same time. Until he retired, he had a reputation as being the best mechanical designer at our facility (out of about 30 or so) and for being extremely capable with the SolidWorks solid modeling program. For reference, when I designed my wind-up powered gyro calibration mechanism last year, the boss I had at the time made the comment that this designer was probably the only other person at the facility who could have done something like that, with one or two others having some chance. Well, on the afternoon of the day of the luncheon, I was invited to the department director's weekly staff meeting (for reference, he's my boss' boss' boss). At the meeting, he had everyone in the room talk about what they are working on and how it's going. I mentioned my progress on my gyro design, and one of the managers on par with my boss piped in, "I was talking to [the designer] about you at the luncheon today, and he told me, 'That kid is a SolidWorks genius!'" I just sort of blinked, stunned, and felt my face turning bright red, "Wow. That is really nice to hear. Thank you for telling me." I thought, but did not say, 'And thank you even more for saying it so blatantly in front of my boss, his boss, and his boss.'
The second program has just started, and I was assigned to it in large part due to my performance on the above-mentioned program. In this program, my job is to first develop the design of the initial test-bed for a new type of gyro, a nuclear magnetic resonance gyro (NMRG), and then miniaturize it. I am one of only 4 people at our facility working on the design, and I am the sole mechanical engineer and doing the entirety of the mechanical design myself. The others on the team are a physicist, an optical engineer, and an electrical engineer.
This would be the continuation of a program that started in the late 60's and ended in 1982, when they had the test gyro down to approximately a 12-inch long by 4-inch diameter cylinder. I say that it would be the continuation of this program except for the fact that all components except for 4 core cells, all drawings, and almost all design and scientific documents about the previous iteration of the program were literally thrown out some time during the last 23 years. Therefore, we are starting nearly from scratch. All we have to go on are a few pictures, a couple of hand-written notebooks with incomplete data and references to no-longer-existing documentation, and the memories of a physicist who is about to retire and who worked on the program in its previous incarnation.
Based on what I know of the requirements, I have the bulk of my conceptual design for the first test bed done, minus selecting and making changes to accommodate for a few commercial-off-the-shelf components. So far, it's about half the size of a Coke can, and should grow to about 3/4 the size of a Coke can by the time the extra components are added. It can accommodate not only the left-over core cells but also new cell designs down to the expected 1mm size without difficulty. We are outsourcing the cell designs, and have a collaboration of different organizations working on them, including UCLA, CalTech, and a company that makes the cells for atomic clocks. After I select these components, I am going to have to develop the coil apparatus that will give us as close as possible to a completely uniform magnetic field throughout the core cell.
Once we prove that we can make a gyro with this technology, it will then become my job to miniaturize it down to 1 cubic centimeter or less. In short, it looks to be difficult as hell and I can say with no sarcasm whatsoever that I am very much looking forward to that stage of the program.
Unfortunately, between work and school, I seem to come home practically brain-dead almost every night. Imagine taking final exams for 8-12 hours, 5-6 days a week for 6 weeks straight, and you will get an idea of how mentally tired I am, and it is starting to really bleed over into physical tiredness. It was already bleeding over the last time Ash_grey_sky saw me, and it has definitely worsened in the last couple of weeks. Thanks to getting ready for the design review on the autonomous robot project, I have had very little free time since Tuesday evening. I had some free time last night, which I spent vegetating in front of the TV for the last half hour of Star Wars Episode III, then playing Guild Wars with Grampablohm and making clear in my conversation with him (through constantly losing my train of thought) just how brain-dead I was feeling.
I think I just proved how tired I am by rambling on about being tired instead of going to sleep and reducing it. On that note, good night.