I'm not sure what this means. Auto shift doesn't work well?
Maybe a better solution would be if I would add a packet to request regular updates.
Maybe something like
struct IS_RAI // Regular AI
{
byte Size; // 8
byte Type; // ISP_RAI
byte ReqI; // 0
byte PLID;
word Interval; // time between updates, 0 to stop
byte Sp2;
byte Sp3;
};I think it should be easy enough though I haven't checked thoroughly yet.
Last edited by Scawen, .
No problem at all, I only wrote a few sentences and briefly looked at the output in my InSim test program.
Good to hear you identified the issue.
At least this one should be easy if you have an atan2 function.
Direction = atan2f(-Vel.x, Vel.y); // in radians
Thanks for the thanks but you don't need to buy an account just to thank me. 
From my point of view it's most likely there could be a bug in your packet reader. For example two packets arriving at once but you only read the first one.
This is a quite fundamental packet so I doubt that it is broken, though I could check tomorrow.
EDIT: I've checked now and I'm getting IS_PLL as expected. I think you should check how many bytes arrived in your TCP read and compare that with the packets you processed. In case a packet is going missing. Packets may arrive more than one in a single read or even be split across multiple reads, which really does happen!
That behaviour, for these inputs that you can hold indefinitely, is documented above but maybe it's not clear enough.
Mentions in the InSim.txt:
// If the Time value is set, that input will return to default after that time.
// This is probably most useful for CS_CHUP / CS_CHDN / CS_FLASH / CS_HORN inputs.
// If you don't use Time then you should send another packet to zero the input.
// CS_CHUP 3 - shift up (set to 1 for a short time then set back to 0)
// CS_CHDN 4 - shift down
I suggest you could use a time like 10 (which is 0.1s). Otherwise (if you use 0 for the Time value) you are just telling it to hold the lever on full time.
EDIT: Tested again and it's still working as expected. Maybe I'll add a bit clearer description in InSim.txt about the need to zero or use the Time value, for those specific controls.
Last edited by Scawen, .
Wow, thanks!
I see it hadn't been updated for a long time.
I've uploaded it to the online system.
https://www.lfs.net/lfs_translate/public_translation_viewer.php
I'm not sure if you would like to do any more of it, but in case you do, I've added your access to the online editing system. I can easily remove your user name from the list if you don't want it there.
Translation page: https://www.lfs.net/lfs_translate/translate.php
You can either click the disc beside each edited line, or edit several lines and click "save whole translation".
I see it hadn't been updated for a long time.
I've uploaded it to the online system.
https://www.lfs.net/lfs_translate/public_translation_viewer.php
I'm not sure if you would like to do any more of it, but in case you do, I've added your access to the online editing system. I can easily remove your user name from the list if you don't want it there.
Translation page: https://www.lfs.net/lfs_translate/translate.php
You can either click the disc beside each edited line, or edit several lines and click "save whole translation".
EDIT: There is a new update below: https://www.lfs.net/forum/post/2114897#post2114897
Here is another update, F4. It's easier to use as you can set PIF_AUTOGEARS and you can set switches and ignition directly instead of toggling them.
Changes from F3 to F4:
The following inputs can now accept values 1:toggle / 2:off / 3:on
CS_IGNITION / CS_EXTRALIGHT / CS_PITSPEED / CS_TCDISABLE / CS_FOGREAR / CS_FOGFRONT
New special value for input 253 sets help flags.
Values that can set set are any combination of the following "Player flags" bits:
PIF_AUTOGEARS // 8 - auto shift
PIF_HELP_B // 64 - brake help
PIF_AUTOCLUTCH // 512 - auto clutch
The default value for an AI is PIF_AUTOCLUTCH only
If you set PIF_AUTOGEARS you don't need to set PIF_AUTOCLUTCH as it is implied
Zero is now an invalid value for CS_MSX
Valid CS_MSX values are 32768 plus or minus 32767 that results in a valid range 1 to 65535
The only possible word value not included is zero and it's better to report as invalid
In this way, an error will be reported if there is any completely zero AIInputVal
Updated section in docs/InSim.txt
Download:
[ link removed - new update available ]
Here is another update, F4. It's easier to use as you can set PIF_AUTOGEARS and you can set switches and ignition directly instead of toggling them.
Changes from F3 to F4:
The following inputs can now accept values 1:toggle / 2:off / 3:on
CS_IGNITION / CS_EXTRALIGHT / CS_PITSPEED / CS_TCDISABLE / CS_FOGREAR / CS_FOGFRONT
New special value for input 253 sets help flags.
Values that can set set are any combination of the following "Player flags" bits:
PIF_AUTOGEARS // 8 - auto shift
PIF_HELP_B // 64 - brake help
PIF_AUTOCLUTCH // 512 - auto clutch
The default value for an AI is PIF_AUTOCLUTCH only
If you set PIF_AUTOGEARS you don't need to set PIF_AUTOCLUTCH as it is implied
Zero is now an invalid value for CS_MSX
Valid CS_MSX values are 32768 plus or minus 32767 that results in a valid range 1 to 65535
The only possible word value not included is zero and it's better to report as invalid
In this way, an error will be reported if there is any completely zero AIInputVal
Updated section in docs/InSim.txt
// AI CONTROL
// ==========
struct AIInputVal // There is an array of these in the AIC packet below
{
byte Input; // Select input value to set
byte Time; // Time to hold (optional, hundredths of a second)
word Value; // Value to set
};
const int AIC_MAX_INPUTS = 20; // NOTE: Increase if CS_NUM is increased
struct IS_AIC // AI Control - variable size
{
byte Size; // 4 + 4 * (number of inputs)
byte Type; // ISP_AIC
byte ReqI; // 0
byte PLID; // Unique ID of AI driver to control
AIInputVal Inputs [AIC_MAX_INPUTS];
};
// If the Time value is set, that input will return to default after that time.
// This is probably most useful for CS_CHUP / CS_CHDN / CS_FLASH / CS_HORN inputs.
// If you don't use Time then you should send another packet to zero the input.
// Values for Input
// CS_MSX 0 - steer: 1 hard left / 32768 centre / 65535 hard right
// CS_THROTTLE 1 - 0 to 65535
// CS_BRAKE 2 - 0 to 65535
// CS_CHUP 3 - shift up (set to 1 for a short time then set back to 0)
// CS_CHDN 4 - shift down
// CS_IGNITION 5 - toggle
// CS_EXTRALIGHT 6 - toggle
// CS_HEADLIGHTS 7 - 1: off / 2: side / 3: low / 4: high
// CS_SIREN 8
// CS_HORN 9
// CS_FLASH 10
// CS_CLUTCH 11 - 0 to 65535
// CS_HANDBRAKE 12 - 0 to 65535
// CS_INDICATORS 13 - 1: cancel / 2: left / 3: right / 4: hazard
// CS_GEAR 14 - for shifter (leave at 255 for sequential control)
// CS_LOOK 15 - 0: none / 4: left / 5: left+ / 6: right / 7: right+
// CS_PITSPEED 16 - toggle
// CS_TCDISABLE 17 - toggle
// CS_FOGREAR 18 - toggle
// CS_FOGFRONT 19 - toggle
// CS_NUM 20 - number of input values above
// The Inputs marked 'toggle' accept the following Values:
// 1 toggle
// 2 switch off
// 3 switch on
// Special values for Input
// 253: set help flags - Value can be any combination of
// PIF_AUTOGEARS // 8 - auto shift
// PIF_HELP_B // 64 - brake help
// PIF_AUTOCLUTCH // 512 - auto clutch
// Default value for an AI driver is PIF_AUTOCLUTCH only
// If you set PIF_AUTOGEARS you don't need to set PIF_AUTOCLUTCH
// 254: reset all inputs
// Most inputs are zero / CS_MSX is 32768 / CS_GEAR is 255
// 255: stop control
// The AI driver will stop the car
// AI car info
// -----------
// Send a SMALL_AII with UVal set to PLID to receive current information about a local car
struct OSMain // included in IS_AII - identical to OutSimMain (main data in OutSim packet)
{
Vector AngVel; // 3 floats, angular velocity vector
float Heading; // anticlockwise from above (Z)
float Pitch; // anticlockwise from right (X)
float Roll; // anticlockwise from front (Y)
Vector Accel; // 3 floats X, Y, Z
Vector Vel; // 3 floats X, Y, Z
Vec Pos; // 3 ints X, Y, Z (1m = 65536)
};
struct IS_AII // AI Info
{
byte Size; // 96
byte Type; // ISP_AII
byte ReqI; // ReqI from the SMALL_AII request packet
byte PLID;
OSMain OSData;
byte Flags; // AIFLAGS_x below
byte Gear; // Reverse:0, Neutral:1, First:2...
byte Sp2;
byte Sp3;
float RPM; // RPM
float SpF0;
float SpF1;
unsigned ShowLights; // Dash lights currently switched on (see DL_x in OutGauge section below)
unsigned SPU1;
unsigned SPU2;
unsigned SPU3;
};
#define AIFLAGS_IGNITION 1 // detect if engine running
//
#define AIFLAGS_CHUP 4 // upshift lever currently held
#define AIFLAGS_CHDN 8 // downshift lever currently held
Download:
[ link removed - new update available ]
Last edited by Scawen, .
I've moved spam and a resolved issue to a cleanup thread.
I request that people who can't understand the benefit of this system should avoid posting here.
I think this can be simply done. I'll probably add an off and on value for the fog and extra light.
EDIT: I think it may be possible to add some help options too:
#define PIF_AUTOGEARS 8
#define PIF_HELP_B 64
#define PIF_AUTOCLUTCH 512
By default an AI has PIF_AUTOCLUTCH set. It may be possible to add an input to set these 3 values optionally.
EDIT2: I did the toggle update this morning and it will be in F4.
The following inputs can now accept values 1:toggle / 2:off / 3:on
CS_IGNITION
CS_EXTRALIGHT
CS_PITSPEED
CS_TCDISABLE
CS_FOGREAR
CS_FOGFRONT
EDIT3: Now I've added the ability to set help flags.
New special value for input 253 sets help flags.
Values that can set set are any combination of the following "Player flags" bits:
PIF_AUTOGEARS // 8 - auto shift
PIF_HELP_B // 64 - brake help
PIF_AUTOCLUTCH // 512 - auto clutch
The default value for an AI is PIF_AUTOCLUTCH only
If you set PIF_AUTOGEARS you don't need to set PIF_AUTOCLUTCH as it is implied
EDIT4: Finally for today, I've made zero an invalid value for CS_MSX
Valid CS_MSX values are 32768 plus or minus 32767 that results in a valid range 1 to 65535
The only possible word value not included is zero and it's better to report as invalid
In this way, an error will be reported if there is any completely zero AIInputVal
I'll update documentation and release F4.
I request that people who can't understand the benefit of this system should avoid posting here.
I think this can be simply done. I'll probably add an off and on value for the fog and extra light.
EDIT: I think it may be possible to add some help options too:
#define PIF_AUTOGEARS 8
#define PIF_HELP_B 64
#define PIF_AUTOCLUTCH 512
By default an AI has PIF_AUTOCLUTCH set. It may be possible to add an input to set these 3 values optionally.
EDIT2: I did the toggle update this morning and it will be in F4.
The following inputs can now accept values 1:toggle / 2:off / 3:on
CS_IGNITION
CS_EXTRALIGHT
CS_PITSPEED
CS_TCDISABLE
CS_FOGREAR
CS_FOGFRONT
EDIT3: Now I've added the ability to set help flags.
New special value for input 253 sets help flags.
Values that can set set are any combination of the following "Player flags" bits:
PIF_AUTOGEARS // 8 - auto shift
PIF_HELP_B // 64 - brake help
PIF_AUTOCLUTCH // 512 - auto clutch
The default value for an AI is PIF_AUTOCLUTCH only
If you set PIF_AUTOGEARS you don't need to set PIF_AUTOCLUTCH as it is implied
EDIT4: Finally for today, I've made zero an invalid value for CS_MSX
Valid CS_MSX values are 32768 plus or minus 32767 that results in a valid range 1 to 65535
The only possible word value not included is zero and it's better to report as invalid
In this way, an error will be reported if there is any completely zero AIInputVal
I'll update documentation and release F4.
Last edited by Scawen, .
I guess the AIInputVal structures you unintentionally submitted as zero were interpreted as:
Input = 0 : CS_MSX (steer)
Time = 0 : stay on
Value = 0 : full left
Input = 0 : CS_MSX (steer)
Time = 0 : stay on
Value = 0 : full left
No floats, the hold times are bytes representing hundredths of a second.
I don't know how that language works but you'd better make sure the byte and word values are combined correctly, resulting in the correct size packet.
That is true.
Checked again and steering does work in F3.
In my test program I have:
Full left lock would be the result of sending zero in the Value field.
In my test program I have:
case 'J':
SendCarPack(car_plid, CS_MSX, 32768 - 2000);
break;
case 'K':
SendCarPack(car_plid, CS_MSX, 32768);
break;
case 'L':
SendCarPack(car_plid, CS_MSX, 32768 + 2000);
break;
Full left lock would be the result of sending zero in the Value field.
Total speculation and in fact complete rubbish.
People are free to suggest what they want and I will choose what I do and how to manage my time.
If other people will be deciding what I do I might as well go and get a job as a programmer and earn 5 times the salary.
I'd be grateful if community members wouldn't try to be self appointed LFS project managers.
I did one afternoon's work and a morning fixing it, on something that people have been waiting for, for longer than people have been waiting for tyre physics. This had no bearing at all on the completion of the tyre physics. Please avoid making up rubbish and allow people to make sensible suggestions if they wish to.
This is the LFS Programmer Forum.
In most cases it is probably best to not comment if you aren't a programmer. Otherwise I would have posted in General LFS Discussion or Test Patch forum.
I don't go and post my opinions in the quantum physics forums because I probably don't know what the **** I am talking about.
In most cases it is probably best to not comment if you aren't a programmer. Otherwise I would have posted in General LFS Discussion or Test Patch forum.
I don't go and post my opinions in the quantum physics forums because I probably don't know what the **** I am talking about.
EDIT: There is a new update below: https://www.lfs.net/forum/post/2114810#post2114810
Here is the update with fixes and improvements for the new system.
I can't get into ray checks as it's not simple enough for this very quick update.
NOTE: If you wrote some code for F2, a small change is required in your AIC packet that has been updated to accept multiple inputs.
Changes from F2 to F3:
InSim:
IS_AIC can now accept multiple input values in a single packet
- a small change must be made if you used IS_AIC from version F2
IS_AIC inputs can now be given a time in hundredths of a second
- after time elapsed the input returns to default value
- this is probably most useful for shift up/down and flash/horn
FIX: Error messages for AIC and AII packets reported as IS_PLH errors
AI:
AI can now spawn in a config with an AI path even if knw has not been generated
FIX: AI now spawn in game with full fuel load if there is no path (or no knw)
FIX: Rare crash in AI code after another car spawned outside path
Here are the updates as seen in docs\InSim.txt
Download: [ link removed, new update available ]
Here is the update with fixes and improvements for the new system.
I can't get into ray checks as it's not simple enough for this very quick update.
NOTE: If you wrote some code for F2, a small change is required in your AIC packet that has been updated to accept multiple inputs.
Changes from F2 to F3:
InSim:
IS_AIC can now accept multiple input values in a single packet
- a small change must be made if you used IS_AIC from version F2
IS_AIC inputs can now be given a time in hundredths of a second
- after time elapsed the input returns to default value
- this is probably most useful for shift up/down and flash/horn
FIX: Error messages for AIC and AII packets reported as IS_PLH errors
AI:
AI can now spawn in a config with an AI path even if knw has not been generated
FIX: AI now spawn in game with full fuel load if there is no path (or no knw)
FIX: Rare crash in AI code after another car spawned outside path
Here are the updates as seen in docs\InSim.txt
// AI CONTROL
// ==========
struct AIInputVal
{
byte Input; // Select input value to set
byte Time; // Time to hold (optional, hundredths of a second)
word Value; // Value to set
};
const int AIC_MAX_INPUTS = 20; // NOTE: Increase if CS_NUM is increased
struct IS_AIC // AI Control - variable size
{
byte Size; // 4 + 4 * (number of inputs)
byte Type; // ISP_AIC
byte ReqI; // 0
byte PLID; // Unique ID of AI driver to control
AIInputVal Inputs [AIC_MAX_INPUTS];
};
// If the Time value is set, that input will return to default after that amount of time.
// This is expected to be most useful for CS_CHUP / CS_CHDN / CS_FLASH / CS_HORN inputs.
// If you don't use Time then you must send another packet to zero the input.
// Values for Input
// CS_MSX 0 - steering : 32768 is centre
// CS_THROTTLE 1 - 0 to 65535
// CS_BRAKE 2 - 0 to 65535
// CS_CHUP 3 - shift up (set to 1 for a short time then set back to 0)
// CS_CHDN 4 - shift down
// CS_IGNITION 5 - set to 1 (auto switch off)
// CS_EXTRALIGHT 6
// CS_HEADLIGHTS 7 - 1: off / 2: side / 3: low / 4: high
// CS_SIREN 8
// CS_HORN 9
// CS_FLASH 10
// CS_CLUTCH 11 - 0 to 65535
// CS_HANDBRAKE 12 - 0 to 65535
// CS_INDICATORS 13 - 1: cancel / 2: left / 3: right / 4: hazard
// CS_GEAR 14 - for shifter (leave at 255 for sequential control)
// CS_LOOK 15 - 0: none / 4: left / 5: left+ / 6: right / 7: right+
// CS_PITSPEED 16
// CS_TCDISABLE 17
// CS_FOGREAR 18
// CS_FOGFRONT 19
// CS_NUM 20 - number of values above
// Special values for Input
// 254 - reset all
// 255 - stop control
// AI car info
// -----------
// Send a SMALL_AII with UVal set to PLID to receive current information about a local car
struct OSMain // included in IS_AII - identical to OutSimMain (main data in OutSim packet)
{
Vector AngVel; // 3 floats, angular velocity vector
float Heading; // anticlockwise from above (Z)
float Pitch; // anticlockwise from right (X)
float Roll; // anticlockwise from front (Y)
Vector Accel; // 3 floats X, Y, Z
Vector Vel; // 3 floats X, Y, Z
Vec Pos; // 3 ints X, Y, Z (1m = 65536)
};
struct IS_AII // AI Info
{
byte Size; // 96
byte Type; // ISP_AII
byte ReqI; // ReqI from the SMALL_AII request packet
byte PLID;
OSMain OSData;
byte Flags; // AIFLAGS_x below
byte Gear; // Reverse:0, Neutral:1, First:2...
byte Sp2;
byte Sp3;
float RPM; // RPM
float SpF0;
float SpF1;
unsigned ShowLights; // Dash lights currently switched on (see DL_x in OutGauge section below)
unsigned SPU1;
unsigned SPU2;
unsigned SPU3;
};
#define AIFLAGS_IGNITION 1 // detect if engine running
//
#define AIFLAGS_CHUP 4 // upshift lever currently held
#define AIFLAGS_CHDN 8 // downshift lever currently held
Download: [ link removed, new update available ]
Last edited by Scawen, .
Reason : fix typo
No, it's not necessary.
But before once again rejecting the mod that was designed to allow some amount of fun, in this case I was able to spend a few hours doing something that allows a huge amount of possibilities. Then rejecting the mod, I'm not just being the nasty developer again.
I believe you are wrong about the calendar and league stuff. A lot of people were interested, especially in the calendar and it lead directly to the league stuff.
After discovering all the bugs and unfinished work in the league stuff, which I encountered while looking at the calendar which was so ugly before, it was near impossible for me to leave the league pages in that state. Victor had basically started a few things there but they were not finished and there were bugs, although basically he has written a good system. Unfortunately for us, he didn't take the time to work through the issues before he got fed up working on LFS. So I'm still left picking up the pieces when I find them.
It's actually a good thing, that I don't like leaving rubbish unfinished code lying around. That's how I work and I believe it leads to good things.
It's pointless to be worried about me delaying some extremely difficult head scratching work that I have been working on this week, to do something that the programmers here can do some super things with.
The tyre stuff is so extremely difficult, requiring a lot of experimentation and even disturbed sleep in the night, trying to think of the next way to attempt to fix the issue I'm trying to deal with. It's not the kind of thing I can just sit there and do it. It required pushes and recovery. I don't really expect people to understand. But after a strong push on Monday to Wednesday it is good for me mentally do to do some rapid coding that can increase the possibilities of LFS. It was 'one day' but there are always a few things to fix. Again, I believe that is a good thing. If I actually knew how long anything would take, I wouldn't even start it.
So I hope some people are very happy now that I have taken the time to explain why I make decisions. It would actually save me time if I could be allowed to do my job without having to explain every step I take. Maybe just watch and see why I do things, then you wouldn't be so confused.
But before once again rejecting the mod that was designed to allow some amount of fun, in this case I was able to spend a few hours doing something that allows a huge amount of possibilities. Then rejecting the mod, I'm not just being the nasty developer again.
I believe you are wrong about the calendar and league stuff. A lot of people were interested, especially in the calendar and it lead directly to the league stuff.
After discovering all the bugs and unfinished work in the league stuff, which I encountered while looking at the calendar which was so ugly before, it was near impossible for me to leave the league pages in that state. Victor had basically started a few things there but they were not finished and there were bugs, although basically he has written a good system. Unfortunately for us, he didn't take the time to work through the issues before he got fed up working on LFS. So I'm still left picking up the pieces when I find them.
It's actually a good thing, that I don't like leaving rubbish unfinished code lying around. That's how I work and I believe it leads to good things.
It's pointless to be worried about me delaying some extremely difficult head scratching work that I have been working on this week, to do something that the programmers here can do some super things with.
The tyre stuff is so extremely difficult, requiring a lot of experimentation and even disturbed sleep in the night, trying to think of the next way to attempt to fix the issue I'm trying to deal with. It's not the kind of thing I can just sit there and do it. It required pushes and recovery. I don't really expect people to understand. But after a strong push on Monday to Wednesday it is good for me mentally do to do some rapid coding that can increase the possibilities of LFS. It was 'one day' but there are always a few things to fix. Again, I believe that is a good thing. If I actually knew how long anything would take, I wouldn't even start it.
So I hope some people are very happy now that I have taken the time to explain why I make decisions. It would actually save me time if I could be allowed to do my job without having to explain every step I take. Maybe just watch and see why I do things, then you wouldn't be so confused.
This appears to be an existing bug, since the AI driving improvements some time ago. It is possible for the crash to appear in the ordinary AI code, if another car has spawned outside the path (so does not have a valid path node). Does that sound a reasonable explanation? I've fixed it but everyone on F will still have the bug. So better not spawn outside the path if ordinary AI are driving around on the track.
Multiple values per packet are done, I plan to add a "Hold time" which could be useful for shift up/down and horn/flash so you can send and forget.
struct AIInputVal
{
byte Input; // Select input value to set
byte Time; // Time to hold (optional)
word Value; // Value to set
};
const int AIC_MAX_INPUTS = 20; // NOTE: Increase if CS_NUM is increased
struct IS_AIC // AI Control - variable size
{
byte Size; // 4 + 4 * (number of inputs)
byte Type; // ISP_AIC
byte ReqI; // 0
byte PLID; // Unique ID of AI driver to control
AIInputVal Inputs [AIC_MAX_INPUTS];
};
254 doesn't switch lights, as zero for those inputs means no change, and they are auto-set to zero after use anyway. I think it is really for all the axes and shift levers, you would have to send a non-zero value to switch off lights.
Thanks for the feedback and bug reports. I've done most of them and will release an update today.
My test seems fine.
- set steering : ok
- set throttle : ok
- send a 254 : steering returns to centre / throttle returns to zero
I might change it to allow multiple values to be set at once in a variable sized IS_AIC (probably would change the zero byte to PLID and then allow multiple 4-byte blocks of "byte Spare/byte Input/word Value"
Yes, I do this all the time when testing at layout square. You set CS_IGNITION to 1
My test seems fine.
- set steering : ok
- set throttle : ok
- send a 254 : steering returns to centre / throttle returns to zero
I might change it to allow multiple values to be set at once in a variable sized IS_AIC (probably would change the zero byte to PLID and then allow multiple 4-byte blocks of "byte Spare/byte Input/word Value"
Yes, I do this all the time when testing at layout square.
You set CS_IGNITION to 1EDIT: A new update is available below: https://www.lfs.net/forum/post/2114731#post2114731
OK, here is the test patch. The only new thing is AI control and data reporting through InSim.
docs/InSim.txt is included with this new section:
Download: [ download removed - F3 is now available ]
OK, here is the test patch. The only new thing is AI control and data reporting through InSim.
docs/InSim.txt is included with this new section:
// AI CONTROL
// ==========
struct IS_AIC // AI Control
{
byte Size; // 8
byte Type; // ISP_AIC
byte ReqI; // 0
byte Zero;
byte PLID; // Unique ID of AI player to control
byte Input; // Select input value to set
word Value; // Value to set
};
// Values for Input
// CS_MSX 0 - steering : 32768 is centre
// CS_THROTTLE 1 - 0 to 65535
// CS_BRAKE 2 - 0 to 65535
// CS_CHUP 3 - shift up (set to 1 for a short time then set back to 0)
// CS_CHDN 4 - shift down
// CS_IGNITION 5 - set to 1 (auto switch off)
// CS_EXTRALIGHT 6
// CS_HEADLIGHTS 7 - 1: off / 2: side / 3: low / 4: high
// CS_SIREN 8
// CS_HORN 9
// CS_FLASH 10
// CS_CLUTCH 11 - 0 to 65535
// CS_HANDBRAKE 12 - 0 to 65535
// CS_INDICATORS 13 - 1: cancel / 2: left / 3: right / 4: hazard
// CS_GEAR 14 - for shifter (leave at 255 for sequential control)
// CS_LOOK 15 - 0: none / 4: left / 5: left+ / 6: right / 7: right+
// CS_PITSPEED 16
// CS_TCDISABLE 17
// CS_FOGREAR 18
// CS_FOGFRONT 19
// CS_NUM 20 - number of values above
// Special values for Input
// 254 - reset all
// 255 - stop control
// AI car info
// -----------
// Send a SMALL_AII with UVal set to PLID to receive current information about a local car
struct OSMain // included in IS_AII - identical to OutSimMain (main data in OutSim packet)
{
Vector AngVel; // 3 floats, angular velocity vector
float Heading; // anticlockwise from above (Z)
float Pitch; // anticlockwise from right (X)
float Roll; // anticlockwise from front (Y)
Vector Accel; // 3 floats X, Y, Z
Vector Vel; // 3 floats X, Y, Z
Vec Pos; // 3 ints X, Y, Z (1m = 65536)
};
struct IS_AII // AI Info
{
byte Size; // 96
byte Type; // ISP_AII
byte ReqI; // ReqI from the SMALL_AII request packet
byte PLID;
OSMain OSData;
byte Flags; // AIFLAGS_x below
byte Gear; // Reverse:0, Neutral:1, First:2...
byte Sp2;
byte Sp3;
float RPM; // RPM
float SpF0;
float SpF1;
unsigned ShowLights; // Dash lights currently switched on (see DL_x in OutGauge section below)
unsigned SPU1;
unsigned SPU2;
unsigned SPU3;
};
#define AIFLAGS_IGNITION 1 // detect if engine running
//
#define AIFLAGS_CHUP 4 // upshift lever currently held
#define AIFLAGS_CHDN 8 // downshift lever currently held
Download: [ download removed - F3 is now available ]
Last edited by Scawen, .
I've replaced that CompCar part with OutSimMain as it's more informative and relevant. I think you need MCI for general awareness.
I don't really want to get into setting intervals for AII packets and you can just request them with a SMALL_AII for now.
I think my testing is done, I can control a car through AIC, even if I haven't tested every possible thing. I seem to be getting info out from AII OK, even if I haven't checked ever single value.
So I'll prepare a simple patch zip.
I don't really want to get into setting intervals for AII packets and you can just request them with a SMALL_AII for now.
I think my testing is done, I can control a car through AIC, even if I haven't tested every possible thing. I seem to be getting info out from AII OK, even if I haven't checked ever single value.
So I'll prepare a simple patch zip.
Yes, CS_MSX is steering, from left to right.
I've been trying to get that car info sorted. What I've got so far is this:
// AI CONTROL
// ----------
struct IS_AIC // AI Control
{
byte Size; // 8
byte Type; // ISP_AIC
byte ReqI; // 0
byte Zero;
byte PLID; // Unique ID of AI player to control
byte Input; // Select input value to set
word Value; // Value to set
};
// Values for Input
// CS_MSX 0 - steering : 32768 is centre
// CS_THROTTLE 1 - 0 to 65535
// CS_BRAKE 2 - 0 to 65535
// CS_CHUP 3 - shift up (set to 1 for a short time then set back to 0)
// CS_CHDN 4 - shift down
// CS_IGNITION 5 - set to 1 (auto switch off)
// CS_EXTRALIGHT 6
// CS_HEADLIGHTS 7 - 1: off / 2: side / 3: low / 4: high
// CS_SIREN 8
// CS_HORN 9
// CS_FLASH 10
// CS_CLUTCH 11 - 0 to 65535
// CS_HANDBRAKE 12 - 0 to 65535
// CS_INDICATORS 13 - 1: cancel / 2: left / 3: right / 4: hazard
// CS_GEAR 14 - for shifter (leave at 255 for sequential control)
// CS_LOOK 15 - 0: none / 4: left / 5: left+ / 6: right / 7: right+
// CS_PITSPEED 16
// CS_TCDISABLE 17
// CS_FOGREAR 18
// CS_FOGFRONT 19
// CS_NUM 20 - number of values above
// Special values for Input
// 254 - reset all
// 255 - stop control
// Send an SMALL_AII with UVal set to PLID to receive current information about a car
struct IS_AII // AI Info
{
byte Size; // 64
byte Type; // ISP_MCI
byte ReqI; // ReqI from the SMALL_AII request packet
byte Zero;
CompCar Info; // As in the MCI packet
// as in outgauge
byte Flags; // AIFLAGS_xelelow
byte Gear; // Reverse:0, Neutral:1, First:2...
byte Sp2;
byte Sp3;
float Spare;
float RPM; // RPM
unsigned ShowLights; // Dash lights (see DL_x in OutGauge section below)
unsigned SPU0;
unsigned SPU1;
unsigned SPU2;
unsigned SPU3;
};
#define AIFLAGS_IGNITION 1 // detect if engine running
//
#define AIFLAGS_CHDN 4 // downshift lever currently held
#define AIFLAGS_CHUP 8 // upshift lever currently held
I've been trying to get that car info sorted. What I've got so far is this:
// AI CONTROL
// ----------
struct IS_AIC // AI Control
{
byte Size; // 8
byte Type; // ISP_AIC
byte ReqI; // 0
byte Zero;
byte PLID; // Unique ID of AI player to control
byte Input; // Select input value to set
word Value; // Value to set
};
// Values for Input
// CS_MSX 0 - steering : 32768 is centre
// CS_THROTTLE 1 - 0 to 65535
// CS_BRAKE 2 - 0 to 65535
// CS_CHUP 3 - shift up (set to 1 for a short time then set back to 0)
// CS_CHDN 4 - shift down
// CS_IGNITION 5 - set to 1 (auto switch off)
// CS_EXTRALIGHT 6
// CS_HEADLIGHTS 7 - 1: off / 2: side / 3: low / 4: high
// CS_SIREN 8
// CS_HORN 9
// CS_FLASH 10
// CS_CLUTCH 11 - 0 to 65535
// CS_HANDBRAKE 12 - 0 to 65535
// CS_INDICATORS 13 - 1: cancel / 2: left / 3: right / 4: hazard
// CS_GEAR 14 - for shifter (leave at 255 for sequential control)
// CS_LOOK 15 - 0: none / 4: left / 5: left+ / 6: right / 7: right+
// CS_PITSPEED 16
// CS_TCDISABLE 17
// CS_FOGREAR 18
// CS_FOGFRONT 19
// CS_NUM 20 - number of values above
// Special values for Input
// 254 - reset all
// 255 - stop control
// Send an SMALL_AII with UVal set to PLID to receive current information about a car
struct IS_AII // AI Info
{
byte Size; // 64
byte Type; // ISP_MCI
byte ReqI; // ReqI from the SMALL_AII request packet
byte Zero;
CompCar Info; // As in the MCI packet
// as in outgauge
byte Flags; // AIFLAGS_xelelow
byte Gear; // Reverse:0, Neutral:1, First:2...
byte Sp2;
byte Sp3;
float Spare;
float RPM; // RPM
unsigned ShowLights; // Dash lights (see DL_x in OutGauge section below)
unsigned SPU0;
unsigned SPU1;
unsigned SPU2;
unsigned SPU3;
};
#define AIFLAGS_IGNITION 1 // detect if engine running
//
#define AIFLAGS_CHDN 4 // downshift lever currently held
#define AIFLAGS_CHUP 8 // upshift lever currently held
I should emphasise that this quick thing will not be good enough for racing. You don't receive detailed or fast enough information for control of a car near the limit. This can only be used for driving cars well within their limits. It's more of a thing for cruise or lessons.
Simple control of AI car
EDIT: A first test is now available below: https://www.lfs.net/forum/post/2114646#post2114646
Hello programmers.
There is a mod waiting for publishing, for the purpose of being "driven" by AI, but all it would actually do is drive along uncontrolled. But I don't really want to support such a bodge mod that bypasses all realism purely so it pulls away all by itself.
But I don't want to prevent something that could be fun so I've come up with a simple idea for just a few hours of work, to allow InSim to control a local car, in a very crude way. I've done a proof of concept and can now drive an AI car from an external program. Actually it disables the AI completely and you just set values, like Steer, throttle, brake, shift, etc.
EDIT: This is only for local AI. It would be far more complicated to do it from the server - that would require new multiplayer packets and I'm not getting into that.
EDIT2: To be clear, it can work for AI that are online, but the InSim must be connected to a local computer that has its own AI on it, and those are the only AI that can be controlled by this new simple packet.
I want to confirm if this could be useful.
It's just like this:
struct IS_AIC // AI Control
{
byte Size; // 8
byte Type; // ISP_AIC
byte ReqI; // 0
byte Zero;
byte PLID; // Unique ID of AI player to control
byte Input; // Select input value to set
word Value; // Value to set
};
// Input values from from CS_NUM in Inputs.h
// Special values
// 254 - reset all
// 255 - stop control
enum
{
CS_MSX,
CS_THROTTLE,
CS_BRAKE,
CS_CHUP,
CS_CHDN,
CS_IGNITION,
CS_EXTRALIGHT,
CS_HEADLIGHTS,
CS_SIREN,
CS_HORN,
CS_FLASH,
CS_CLUTCH,
CS_HANDBRAKE,
CS_INDICATORS,
CS_GEAR,
CS_LOOK,
CS_PITSPEED,
CS_TCDISABLE,
CS_FOGREAR,
CS_FOGFRONT,
CS_NUM
};
I think you might need some minimal feedback from the car, via a request info packet, and it could tell you some things like the lights and ignition status. I suppose you can get speed / position, etc from other packets like MCI? Or maybe you need something like an MCI for one car, which you can receive on request for a local AI, and maybe it has a little more date in it?
I want to keep it simple as I don't want this to go on more than this afternoon as I have other things that I am working on.
Hello programmers.
There is a mod waiting for publishing, for the purpose of being "driven" by AI, but all it would actually do is drive along uncontrolled. But I don't really want to support such a bodge mod that bypasses all realism purely so it pulls away all by itself.
But I don't want to prevent something that could be fun so I've come up with a simple idea for just a few hours of work, to allow InSim to control a local car, in a very crude way. I've done a proof of concept and can now drive an AI car from an external program. Actually it disables the AI completely and you just set values, like Steer, throttle, brake, shift, etc.
EDIT: This is only for local AI. It would be far more complicated to do it from the server - that would require new multiplayer packets and I'm not getting into that.
EDIT2: To be clear, it can work for AI that are online, but the InSim must be connected to a local computer that has its own AI on it, and those are the only AI that can be controlled by this new simple packet.
I want to confirm if this could be useful.
It's just like this:
struct IS_AIC // AI Control
{
byte Size; // 8
byte Type; // ISP_AIC
byte ReqI; // 0
byte Zero;
byte PLID; // Unique ID of AI player to control
byte Input; // Select input value to set
word Value; // Value to set
};
// Input values from from CS_NUM in Inputs.h
// Special values
// 254 - reset all
// 255 - stop control
enum
{
CS_MSX,
CS_THROTTLE,
CS_BRAKE,
CS_CHUP,
CS_CHDN,
CS_IGNITION,
CS_EXTRALIGHT,
CS_HEADLIGHTS,
CS_SIREN,
CS_HORN,
CS_FLASH,
CS_CLUTCH,
CS_HANDBRAKE,
CS_INDICATORS,
CS_GEAR,
CS_LOOK,
CS_PITSPEED,
CS_TCDISABLE,
CS_FOGREAR,
CS_FOGFRONT,
CS_NUM
};
I think you might need some minimal feedback from the car, via a request info packet, and it could tell you some things like the lights and ignition status. I suppose you can get speed / position, etc from other packets like MCI? Or maybe you need something like an MCI for one car, which you can receive on request for a local AI, and maybe it has a little more date in it?
I want to keep it simple as I don't want this to go on more than this afternoon as I have other things that I am working on.
Last edited by Scawen, .
I have done some research today. Based on income data I was able to move Argentina and Chile from the middle price band into the lower price band.
As I understand it both of these South American countries have had some economic turmoil in the past 8 years that have affected the average income.
As I understand it both of these South American countries have had some economic turmoil in the past 8 years that have affected the average income.
Maybe you can post on that thread that Sobis linked to?
Hopefully a server admin will reply to you there.
Hopefully a server admin will reply to you there.
I should add that this can be done more cleanly (no spectating of players that have already joined). Instead, an InSim program can allow them to spawn or not depending on their selected car.
- Enable join request with ISF_REQ_JOIN
- Use the IS_SLC packet to detect which car has been selected
- Respond to join requests with IS_JRR packets (JRR_REJECT or JRR_SPAWN)
See InSim.txt
- Enable join request with ISF_REQ_JOIN
- Use the IS_SLC packet to detect which car has been selected
- Respond to join requests with IS_JRR packets (JRR_REJECT or JRR_SPAWN)
See InSim.txt
That's because...
1) The flawed model is simple to code
2) It is actually 'logical' (although not physical)
I'm interested to try this, even the 'fake' version (making throttle response more realistic without doing a full engine simulation).
My understanding of the physical reason for this is that the diameter of the fully open butterfly valve is chosen to allow sufficient airflow at the engine's maximum power rpm. At low rpm the engine draws in air at a much slower rate so there isn't much resistance and so at low rpm it makes barely any difference whether the butterfly valves are fully open or half open. At low rpm when air flow is low, what makes the difference is whether the valves are a tiny bit open or slightly more open.
These days I'm sure it is more complicated as there is throttle mapping. So there is not a linear mapping between pedal and rotation of the butterfly valves. I don't know yet how the throttle mapping might be varied at different rpm. I'd be interested in that (but it doesn't affect the ability to try a test).
Also to note, not exactly the subject of this thread but even in a traditional engine without throttle mapping, and even at max power rpm, our model is also wrong, because the throttle response is linear, but that doesn't match the characteristic of butterfly valves which produce an S-shaped flow-vs-resistance characteristic. I'm sure that S-shape would be straightened by the throttle mapping (but just to repeat, I know that is not really the topic of this thread).
I do have higher priorities at the moment though, like trying to get to a state where we can do that release. This post is more to acknowledge that I'm interested in this subject, for those who didn't know that already.
FGED GREDG RDFGDR GSFDG