Aston Sim Engineer / F1 Sim job
#1
Rennlist Hoonigan
which cost no drachmas
Lifetime Rennlist
Member
Rennlist
Site Sponsor
which cost no drachmas
Lifetime Rennlist
Member
Rennlist
Site Sponsor
Thread Starter
![Default](https://rennlist.com/forums/images/icons/icon1.gif)
Maybe better suited for the Sim Forum, I thought folks here might find the job description and requirements interesting https://www.motorsportjobs.com/en/jo...80cp_JA.Vbb5eh
The requirements around Matlab and C/C++ for programming are interesting along with relevant request of motion cueing/perception, HIL Sim, and LiDar, A bit out of my leage!
The requirements around Matlab and C/C++ for programming are interesting along with relevant request of motion cueing/perception, HIL Sim, and LiDar, A bit out of my leage!
#2
Rennlist Member
![Default](https://rennlist.com/forums/images/icons/icon1.gif)
My job is not so different, but also very different.
I work in controls and industrial automation. So, I do a lot of modeling, programming, and process control. I’ve dabbled in games a little in a hobby capacity, and it’s actually similar to open loop controls (no feedback). In controls, we want to start with a model and calculate the required output to produce a result as close as possible. This is the part that’s similar. But then, we (typically) want to close a feedback loop and adjust the output based on error. But we want to make that error as small as possible by making the initial model as close as possible.
A simple example is spinning an eccentric load in a vertical circle at a constant speed with an async AC motor. You start with a constant frequency output. But the load slows on the way up and overruns on the way down. So, you add an encoder for position feedback to close the speed control loop. Now, the control reacts to speed error so as it slows on the way up, the output frequency increases and when it overruns on the way down, the output frequency decreases. But it has to have error to react to. Better yet would be to reduce the initial error by tracking the position, calculating the required torque (which will look like a sine wave for a circular motion) and adding a torque feed forward to the control loop. So, you tell the drive to send more current ahead of time on the way up and back off on the way down. Now, the speed error is greatly reduced and the speed control loop has a much easier job. One of the things this means is you can tune the speed control loop hotter because the overall amount of error is smaller, it can react quicker, and stay even closer to the target profile.
I’m new to HPDE (first track weekend last month), but I’m really enjoying viewing driving on the track from a control system perspective. For example, I see picking your line as part of the initial modeling and the various feedback information (G, visual, etc) as closing the driver control loop. And just like the motor example, the better the initial modeling, the less error to correct and the closer to the limit to the driver can maintain.
It’s hard to tune a tight control if the initial model is way off. At the other end, if the model was perfect*, there would be no error to correct.
Being new, I have a poor model and can’t stay close to limit because of the amount of error I have to manage. If that makes sense.
But my Apex Pro arrived today, so I’m hoping to close the loop on my G force next track weekend, which is in 10 days for me. 😃
To tie it back into the simulator, I see it as the modeling portion, without feedback. They may have “simulated feedback”, but I just see that as more modeling.
As it may relate to sim driving (which I’d like to get into at the end of the year), I see the driver as the top control loop in the system, taking in whatever feedback is available and outputting commands to the car (or sim).
As an aside, I’m finding the logging and data aspect super interesting. Looking at trends and mentally walking up and down derivatives and integrals, through multiple cascading control loops, is just day job, which I do because I enjoy. I just lack a lot of domain knowledge as it applies to the track, so I’m slowly working on that.
Anyway, just my perspective.
*perfect is not possible; or at least not practical; hence the feedback to close the loop
I work in controls and industrial automation. So, I do a lot of modeling, programming, and process control. I’ve dabbled in games a little in a hobby capacity, and it’s actually similar to open loop controls (no feedback). In controls, we want to start with a model and calculate the required output to produce a result as close as possible. This is the part that’s similar. But then, we (typically) want to close a feedback loop and adjust the output based on error. But we want to make that error as small as possible by making the initial model as close as possible.
A simple example is spinning an eccentric load in a vertical circle at a constant speed with an async AC motor. You start with a constant frequency output. But the load slows on the way up and overruns on the way down. So, you add an encoder for position feedback to close the speed control loop. Now, the control reacts to speed error so as it slows on the way up, the output frequency increases and when it overruns on the way down, the output frequency decreases. But it has to have error to react to. Better yet would be to reduce the initial error by tracking the position, calculating the required torque (which will look like a sine wave for a circular motion) and adding a torque feed forward to the control loop. So, you tell the drive to send more current ahead of time on the way up and back off on the way down. Now, the speed error is greatly reduced and the speed control loop has a much easier job. One of the things this means is you can tune the speed control loop hotter because the overall amount of error is smaller, it can react quicker, and stay even closer to the target profile.
I’m new to HPDE (first track weekend last month), but I’m really enjoying viewing driving on the track from a control system perspective. For example, I see picking your line as part of the initial modeling and the various feedback information (G, visual, etc) as closing the driver control loop. And just like the motor example, the better the initial modeling, the less error to correct and the closer to the limit to the driver can maintain.
It’s hard to tune a tight control if the initial model is way off. At the other end, if the model was perfect*, there would be no error to correct.
Being new, I have a poor model and can’t stay close to limit because of the amount of error I have to manage. If that makes sense.
But my Apex Pro arrived today, so I’m hoping to close the loop on my G force next track weekend, which is in 10 days for me. 😃
To tie it back into the simulator, I see it as the modeling portion, without feedback. They may have “simulated feedback”, but I just see that as more modeling.
As it may relate to sim driving (which I’d like to get into at the end of the year), I see the driver as the top control loop in the system, taking in whatever feedback is available and outputting commands to the car (or sim).
As an aside, I’m finding the logging and data aspect super interesting. Looking at trends and mentally walking up and down derivatives and integrals, through multiple cascading control loops, is just day job, which I do because I enjoy. I just lack a lot of domain knowledge as it applies to the track, so I’m slowly working on that.
Anyway, just my perspective.
*perfect is not possible; or at least not practical; hence the feedback to close the loop
The following 3 users liked this post by jhenson29:
#3
Rennlist Member
Join Date: Apr 2017
Location: The Swamps of Jersey/WGI/VIR...
Posts: 6,509
Received 1,683 Likes
on
1,143 Posts
![Default](https://rennlist.com/forums/images/icons/icon1.gif)
Matt, you know how much I love the data thing, so I sent in my resume and used you as a reference…
The following 2 users liked this post by dgrobs:
Matt Romanowski (06-13-2024),
ProCoach (06-13-2024)
#4
Rennlist
Basic Site Sponsor
Basic Site Sponsor
Join Date: Jun 2008
Location: Durham, NC and Virginia International Raceway
Posts: 19,069
Received 3,212 Likes
on
1,851 Posts
![Default](https://rennlist.com/forums/images/icons/icon1.gif)
![hiha](https://rennlist.com/forums/graemlins/roflmao.gif)
One can dream!
__________________
-Peter Krause
www.peterkrause.net
www.gofasternow.com
"Combining the Art and Science of Driving Fast!"
Specializing in Professional, Private Driver Performance Evaluation and Optimization
Consultation Available Remotely and at VIRginia International Raceway
-Peter Krause
www.peterkrause.net
www.gofasternow.com
"Combining the Art and Science of Driving Fast!"
Specializing in Professional, Private Driver Performance Evaluation and Optimization
Consultation Available Remotely and at VIRginia International Raceway
#5
Rennlist Member
Join Date: Apr 2017
Location: The Swamps of Jersey/WGI/VIR...
Posts: 6,509
Received 1,683 Likes
on
1,143 Posts
The following users liked this post:
ProCoach (06-13-2024)
The following users liked this post:
ProCoach (06-13-2024)
#7
Rennlist Member
Join Date: Apr 2017
Location: The Swamps of Jersey/WGI/VIR...
Posts: 6,509
Received 1,683 Likes
on
1,143 Posts
The following users liked this post:
ProCoach (06-13-2024)