Powerdrome - Revision history

Dave Mark: /* Multi-tier planning architecture through virtual yoke */ formatted the list of behaviors

2011-07-23T17:41:08Z

Multi-tier planning architecture through virtual yoke: formatted the list of behaviors

Dave Mark: Fixed author stamps

2011-07-22T00:49:50Z

Fixed author stamps

Brett@tilda.plus.com at 23:48, 21 July 2011

2011-07-21T23:48:47Z

Dave Mark: Initial creation for Brett

2011-07-20T00:55:15Z

Initial creation for Brett

New page

{{template:HistoryOfAI}}

{{GameInfoBox
|gametitle=Powerdrom
|imagefile=PowerdromeBox.jpg
|developer=Argonaut Software
|publisher=Mad Duck Productions
|year= 2004
|platforms=PS2, XBox, PC
|genres=Racing
|era=Tactical Reasoning
}}

Powerdrome is a futuristic racing game where players race jet-engined, anti-gravity bikes called 'blades' around closed tracks.

==Description of AI Behavior==
(What kinds of things was the AI supposed to do?)

===Notable Behaviors===
(What really special things did the AI have to perform that made it stand out?)

==Architectures==

===Each Arch. Should Have a Heading Like This===

{{AuthorStamp|name=Your Name|url=Your Name}}

===Another Architecture===

==Post-Mortem==

===What Worked===

====Individual Good Thing====

Something really went well.

====Individual Good Thing====

Something really went well.

===What Didn't Work===

====Individual Bad Thing====
Something really sucked.

====Individual Bad Thing====
Something really sucked.

===Lessons Learned===

==Reception by Public==
(Was there positive or negative reception from the players and/or media?)

==References==

==External Links==
(ADD LATER)

@@ Line 31: / Line 31: @@
 {{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Divided into a number of feed forward sections: a line following layer, a personality layer, the A.I. passed signals down through a stack based virtual yoke.  This generated (predominantly) gas and steering signals similar to the player.
-Noted here is the personality layer, which based on blades around it - and the relationship to those blades - could chose to affect the signals of the line following layer with extra steering - towards blade as it came level => side ram (aggressive), -ve ram => getting out of the way (friendly), sidewards bias with a blade in front => (overtaking), overdriven acceleration to a blade in front => forward ram (aggressive), artificially deceleration to a blade behind => blocking (aggressive).
+Noted here is the personality layer which was based on blades around it&mdash;and the relationship to those blades&mdash;could chose to affect the signals of the line following layer with extra steering:
 The use of boost was also intended to be plugged into personality with the idea that some characters would just like to get ahead, rather than be conservative.

@@ Line 17: / Line 17: @@
 ==Description of AI Behavior==
 Being a pure racing simulation, the A.I. attempted to provide a competative race on a multitude of tracks often with multiple routes.  It tried to do this reflecting the personality of each of the relevant characters
@@ Line 24: / Line 23: @@
 ==Architectures==
 ===Player assisted learning===
-Various people in the company drove a number of laps around various tracks recording telemetry.  The best were chosen primarily to provide appropriate attack velocities and lines for things like corners.  A combination of the multiple routes were driven.
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Various people in the company drove a number of laps around various tracks recording telemetry.  The best were chosen primarily to provide appropriate attack velocities and lines for things like corners.  A combination of the multiple routes were driven.
 For A.I. training, the single player time trial mode was used.  The A.I. continually aimed to beat the best time, driving the best line as closely as possible and dropping telemetry as it went.  If it set a new time, this became that which it tried to achieved.  The natural curvature of the improved driving line therefore started to reflect what the A.I., for that particular blade, could drive inside it's own heuristics.
 ===Multi-tier planning architecture through virtual yoke===
-Divided into a number of feed forward sections: a line following layer, a personality layer, the A.I. passed signals down through a stack based virtual yoke.  This generated (predominantly) gas and steering signals similar to the player.
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Divided into a number of feed forward sections: a line following layer, a personality layer, the A.I. passed signals down through a stack based virtual yoke.  This generated (predominantly) gas and steering signals similar to the player.
 Noted here is the personality layer, which based on blades around it - and the relationship to those blades - could chose to affect the signals of the line following layer with extra steering - towards blade as it came level => side ram (aggressive), -ve ram => getting out of the way (friendly), sidewards bias with a blade in front => (overtaking), overdriven acceleration to a blade in front => forward ram (aggressive), artificially deceleration to a blade behind => blocking (aggressive).
@@ Line 38: / Line 36: @@
 ===Inverse kinematics===
-The A.I. attempts to accelerate/decelerate to acquire the given best line speed * some slightly increased scaler.  It tries to acquire the speed at the next time step, clamped by an acceptable acceleration [worked out by engine max F/m].  Because they fly under the same mechanics as the player, this usually means the players speed is reasonably attainable.  By using a scaler to slightly increase target values so that it is never quite, a smoothed continuous signal is obtained.
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}The A.I. attempts to accelerate/decelerate to acquire the given best line speed * some slightly increased scaler.  It tries to acquire the speed at the next time step, clamped by an acceptable acceleration [worked out by engine max F/m].  Because they fly under the same mechanics as the player, this usually means the players speed is reasonably attainable.  By using a scaler to slightly increase target values so that it is never quite, a smoothed continuous signal is obtained.
-Steering was originally based on a fan approach generated off line for each blade.  Various degrees of steer were recorded under different conditions, the physics being integrated over a set time window to work out the relative translation from forward direction, expected in the future.  This would form part of an n directional table where various key factors, side slip, speed bracket were weighted into the equation.
+Steering was originally based on a fan approach generated off line for each blade.  Various degrees of steer were recorded under different conditions, the physics being integrated over a set time window to work out the relative translation from forward direction, expected in the future.  This would form part of an ''n''-directional table where various key factors, side slip, speed bracket were weighted into the equation.
 As the data grew, the effort was somewhat scaled back to just be hard left and hard right steering.  Then a given run time state of blade, would provide the relevant hard left, hard right table entry and linear interpolation from -1 being the left most point to +1 being the right determined the best steering value.
 ===Obstacle avoidance===
-The OOBB of known obstacles were translated into 2D (horizontal occupation + distance) track space and this used to affect steering values.  By mini-maxing the horizontal occupation, obstacle clusters could be avoided [Laming, AI Wisdom 2]
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}The OOBB of known obstacles were translated into 2D (horizontal occupation + distance) track space and this used to affect steering values.  By mini-maxing the horizontal occupation, obstacle clusters could be avoided [Laming, AI Wisdom 2]
 ==Post-Mortem==
 ===What Worked===
@@ Line 58: / Line 55: @@
 ====Gesturing in replays====
-Although gestures were partly driven by race time, by the personality layer determining a move, most of it was actually a cheat - added instead during replay.  Depending on the replay telemetry, we increased the number of gestures at key times (after collisions - accidental or not), on entering 1st place etc.. and provided the rules were reasonable it looked like they all had more personality than they really did
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Although gestures were partly driven by race time, by the personality layer determining a move, most of it was actually a cheat - added instead during replay.  Depending on the replay telemetry, we increased the number of gestures at key times (after collisions - accidental or not), on entering 1st place etc.. and provided the rules were reasonable it looked like they all had more personality than they really did
 ====Debugging====
-Graphical debugging of avoidance, track and steering feathers.  The ability to pause the simulation (but not camera) and look around.  [Laming, AI Wisdom 2]
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Graphical debugging of avoidance, track and steering feathers.  The ability to pause the simulation (but not camera) and look around.  [Laming, AI Wisdom 2]
 ===What Didn't Work===
 ====Steering by table====
-Fitting steering curves by a multidimensional table of dependent variable brackets.  Reallistically there were too many parameters and we were relying on instinct to tell us which ones to base things on.  The table quickly went n-dimensional.  These days, you'd alter at runtime using a PID controller
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Fitting steering curves by a multidimensional table of dependent variable brackets.  Reallistically there were too many parameters and we were relying on instinct to tell us which ones to base things on.  The table quickly went n-dimensional.  These days, you'd alter at runtime using a PID controller
 ====Imparting personality at 1200km/h====
-Trying to co-ordinate aggressive movements at speeds approaching 1200km/h is difficult.  Often rams or swipes would miss, or at most be ineffectual.
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Trying to co-ordinate aggressive movements at speeds approaching 1200km/h is difficult.  Often rams or swipes would miss, or at most be ineffectual.
 ====Artistically fitted tracks vs physical speed====
-The artists treated both track and blade as a traditional racing circuit.  But when you're flying a tonne of metal without friction at 1200km/h, bends need a huge track width.  Because both player and A.I. often couldn't make them, the side force field profile needed to be ridiculously ramped up - taking away some of the skill
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}The artists treated both track and blade as a traditional racing circuit.  But when you're flying a tonne of metal without friction at 1200km/h, bends need a huge track width.  Because both player and A.I. often couldn't make them, the side force field profile needed to be ridiculously ramped up - taking away some of the skill
 ===Lessons Learned===
-Having a realistic race, isn't necessary the best race.  We probably should have used light rubber bandedness to keep blades together more.
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}Having a realistic race, isn't necessary the best race.  We probably should have used light rubber bandedness to keep blades together more.
 ==Reception by Public==
-The game received mixed reviews and a 69% Metacritic.  To our knowledge the A.I. was never accused of being too stupid or obviously cheating, passing the A.I. was just there test.
+{{AuthorStamp|name=Brett Laming|url=http://www.linkedin.com/in/brettlaming}}The game received mixed reviews and a 69% Metacritic.  To our knowledge the A.I. was never accused of being too stupid or obviously cheating, passing the A.I. was just there test.
 ==References==

@@ Line 17: / Line 17: @@
 ==Description of AI Behavior==
-(What kinds of things was the AI supposed to do?)
+{{AuthorStamp|name=Brett Laming|url=Your Name}}
+Being a pure racing simulation, the A.I. attempted to provide a competative race on a multitude of tracks often with multiple routes.  It tried to do this reflecting the personality of each of the relevant characters
 ===Notable Behaviors===
-(What really special things did the AI have to perform that made it stand out?)
+Notable behaviours including some degree of gesturing based on personality and move execution.
-==Architectures==
+===Multi-tier planning architecture through virtual yoke===
-===Each Arch. Should Have a Heading Like This===
+Noted here is the personality layer, which based on blades around it - and the relationship to those blades - could chose to affect the signals of the line following layer with extra steering - towards blade as it came level => side ram (aggressive), -ve ram => getting out of the way (friendly), sidewards bias with a blade in front => (overtaking), overdriven acceleration to a blade in front => forward ram (aggressive), artificially deceleration to a blade behind => blocking (aggressive).
-<!-- If you are filling this out based on first-hand knowledge
-fill out the AuthorStamp below. Else delete it. -->
-{{AuthorStamp|name=Your Name|url=Your Name}}
-===Another Architecture===
+The use of boost was also intended to be plugged into personality with the idea that some characters would just like to get ahead, rather than be conservative.
 ==Post-Mortem==
 ===What Worked===
-====Individual Good Thing====
+====Player assisted learning====
-Something really went well.
+Really allowed some nice player quirks to come out, tempered by an A.I. that then refined your lap.
-====Individual Good Thing====
+====Gesturing in replays====
-Something really went well.
+Although gestures were partly driven by race time, by the personality layer determining a move, most of it was actually a cheat - added instead during replay.  Depending on the replay telemetry, we increased the number of gestures at key times (after collisions - accidental or not), on entering 1st place etc.. and provided the rules were reasonable it looked like they all had more personality than they really did
 ===What Didn't Work===
-====Individual Bad Thing====
+====Steering by table====
-Something really sucked.
+Fitting steering curves by a multidimensional table of dependent variable brackets.  Reallistically there were too many parameters and we were relying on instinct to tell us which ones to base things on.  The table quickly went n-dimensional.  These days, you'd alter at runtime using a PID controller
-====Individual Bad Thing====
+====Imparting personality at 1200km/h====
-Something really sucked.
+Trying to co-ordinate aggressive movements at speeds approaching 1200km/h is difficult.  Often rams or swipes would miss, or at most be ineffectual.
 ===Lessons Learned===
+Having a realistic race, isn't necessary the best race.  We probably should have used light rubber bandedness to keep blades together more.
-−
 ==Reception by Public==
-(Was there positive or negative reception from the players and/or media?)
+The game received mixed reviews and a 69% Metacritic.  To our knowledge the A.I. was never accused of being too stupid or obviously cheating, passing the A.I. was just there test.
-−
-−
 ==References==
+A.I. Wisdom 2 - todo.
 ==External Links==