Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 2 3 [4] 5 6 ... 18

Author Topic: WarBiz Arms Race // Victory Services Thread (Turn 5 Revision Phase)  (Read 16754 times)

Jerick

  • Bay Watcher
    • View Profile

Quote from: Votebox
Heavy Assault Gear
Scourger revision project (1): frostgiant
Basic Overshield Generator - Secondary Power Core (2): frostgiant, Happerry
Bane Heavy Autolaser (1): Madman

Light Assault Gear
Basic Jump Pack - Variable Nozzle Array (2): Happerry, Jerick

Engineer Gear]
Illuminator Target Illumination Device (1): Madman
Autotarget Illuminator Autobrain(1): Jerick
My reasoning for my choices; adding some cameras and thermal imaging sensors with some software strikes me as much easier than thermal imaging sensors+ changing how the beam fires +making the beam hit in a cross pattern on the target consistently (something that wouldn't work too well if the surface hit was changing angle relative to the turret). As for the jump pack nozzle array it's something that might be useful scaled up for our VTOL craft which I beleive we were planning on.
Logged

Madman198237

  • Bay Watcher
    • View Profile

Jump pack nozzles are very different from VTOL thrusters (which we already have a far-superior example of on our present VTOL craft).

My reasoning is that pattern recognition is incredibly difficult for computers, so why attempt pattern recognition when it's much easier to stick a series of small lasers (we already have huge amounts of laser tech) in an x shape and have the computer reliably pick out a series of dots on a target.
Logged
We shall make the highest quality of quality quantities of soldiers with quantities of quality.

Jerick

  • Bay Watcher
    • View Profile

I apolgise I misread your proposal, I was under the impression the x blast was coming from the turret but my point still stands; your way is more difficult. In fact it means our engnieers would need some considerably larger and bulkier gear to do this, on par with having a whole other weapon. In my estimation it'd have to be at least the size of an SMG if they are to heat enemy targets at range enough for the turret to lock on while in the middle of a fire fight.

As for the computer tracking you're vastly overestimating how difficult it is. To give you the broad strokes how a system like this would work is that it would find groups of pixels that are consistent across multiple frames in colour and relative grouping. It'll naturally take a few frames to identify one of these groups of pixels which from here on out I will refer to as a feature. Background features don't change much and often can be filtered out because of that. Now when something moves across the camera's view it will do so as a moving set of features. Some features will disappear and new ones will be identified. But even as the number of features fluctuate as the object moves there will be enough consistent features to give the computer a high confidence of the object's position and speed. Now if you say add in more cameras then things start to get fun. You can compare the features it sees with the features the other cameras see and don't see. Using multiple cameras of known positions you can plot the features they see in three dimensional space and start to build a 3d model of the moving object. With the features tied to coordinates on a 3d model it can track changes in orientation. We just need the tracking of the features thing on our turret though, we don't need the gun to figure out 3d models of the targets.

I object most strongly to the idea that the future, interplanetary, death is an inconvenience, nanite wielding space mercs would have a problem with tech that in the real world we have been using for decades.

Logged

Madman198237

  • Bay Watcher
    • View Profile

...If it was that simple and that fast AI-driven robots would most likely be far more capable than they actually are. The problem with tracking pixels is that a pixel is an angular measurement. If the object is moving towards or away from you any identifiable "features" would change in size rapidly, and if the object were moving side-to-side as well then the features would also be changing in shape. Yes, eventually your system would indeed have time to pick out that that thing moving is in fact an enemy tank, but it would have to stare at it for long enough that the tank crew will be able to take a coffee break, lunch break, and then a tea break before killing all your allies.

We have combat-usable lasers on par with rifle rounds, compared to those what I'm asking for is a glorified laser pointer.
Logged
We shall make the highest quality of quality quantities of soldiers with quantities of quality.

dgr11897

  • Bay Watcher
    • View Profile

Quote from: Votebox
Heavy Assault Gear
Scourger revision project (1): frostgiant
Basic Overshield Generator - Secondary Power Core (3): frostgiant, Happerry, DGR
Bane Heavy Autolaser (1): Madman

Light Assault Gear
Basic Jump Pack - Variable Nozzle Array (3): Happerry, Jerick, DGR

Engineer Gear]
Illuminator Target Illumination Device (1): Madman
Autotarget Illuminator Autobrain(1): Jerick
Logged
Quote from: evicted Saint on discord
Weaponizing Jesus isn't something you do turn 4
Quote from: Alice on a different discord, to iridium, kind of.
hold on, let me keep blowing kisses at him until he stops
My Power armor arms race

Jerick

  • Bay Watcher
    • View Profile

Quote
...If it was that simple and that fast AI-driven robots would most likely be far more capable than they actually are.
I suspect you are confused. That is the only way this response makes any sense. I think you might be conflating motion tracking with image recognition? To be perfectly clear what I described does not identify the object at any point. It simply spots that something is moving and gives you the position and it's speed. Image recognition is something that is completely different and is the thing that has implications for AI because it requires machine learning. Because that is where a computer program identifies the contents of an image using only the image and a trained neural net. Motion tracking is where the computer examines two separate images for the differences between them. There is no machine learning involved, it is simple math and has no more relevance to AI development than an ultrasound motion detector has.

Quote
The problem with tracking pixels is that a pixel is an angular measurement. If the object is moving towards or away from you any identifiable "features" would change in size rapidly, and if the object were moving side-to-side as well then the features would also be changing in shape.
I said I would give you the broad strokes of something that is relatively simple that would work and I did, if you wish me to go into more detail then I shall; features are lost and gained rapidly the program tracks the overall cluster of features rather than any specific ones. It is in effect tracking changes. I perhaps could have made that clearer in my last post. It turning side to side, coming closer or retreating far away will not diminish the programs ability to give you its position and speed only it vanishing, becoming still enough to filtered out as the background (good filtering code will stop this) or becoming indistinguishable from the background will stop its tracking. It needs to stop having features distinguishable from the background for the lock to be broken, it doesn't matter what those features are. Now for tracking a feature's distance from the camera there is a couple of ways of doing it. The best way is to have multiple cameras because you can math the changes in what they see compared to each other.

Quote
Yes, eventually your system would indeed have time to pick out that that thing moving is in fact an enemy tank, but it would have to stare at it for long enough that the tank crew will be able to take a coffee break, lunch break, and then a tea break before killing all your allies.
What I described would never pick out that that thing moving is an enemy tank because that is not what it does. It does not select targets the engineer does, all it does is keep shoot the thing the engineer pointed it at regardless which way it moves. That's all. It tracks movement. It doesn't identify. It tracks movement. It could probably do so pretty quickly too, heck most phones these days can motion track quite effectively and I refuse the idea that Victory Services would have trouble putting a processor of equal power into military equipment.

Quote
We have combat-usable lasers on par with rifle rounds, compared to those what I'm asking for is a glorified laser pointer.
Legitimate question here; if you're suggesting a glorified laser pointer what's wrong with an actual laser pointer? It's much closer to the laser designator already in the design and there's nothing big difficult or awkward about it. Actually why not keep the current designator and just have it track the heat of its impact point after the initial firing?

Edit:Hrrm reread the autobrain proposal, that makes note of hitting known weakpoints doesn't it? That still probably won't be as difficult as madman is making out but some how I was under the impression it was just an autotracker. Apologies for the miscommunication Madman it was me who was confused.
« Last Edit: May 21, 2019, 06:45:00 pm by Jerick »
Logged

dgr11897

  • Bay Watcher
    • View Profile

Quote from: Illuminator retinal based targeting
Rather than futz around with sensors and target detection, or continuing to go with a somewhat flawed designation system. This solution to the illuminator's targeting issues feeds telemetry from a helmet with some gimbal mounted rangefinders, a compass/relative orientation sensor and a GPS chip to allow the turret to derive roughly where to aim. The rangefinders are aimed by the users retinas, information on where they are looking being derived from cameras inside the helmet and off the shelf retinal/eye tracking software installed on a small computer located on the rest of the apparatus. Which is composed of a backpack carried computer and a controller with two buttons, arm/disarm, and fire. The system works as follows, the operator arms the system, looks at a target, the system displays the turret's current aimpoint (possibly also giving the turret a visible laser pointer to indicate when it's on target), then the operator presses a button to fire the device once the aimpoint is locked onto the target. Importantly the helmet is feeding range and position telemetry after deriving where to look from the retinal tracking system inside the target, letting the system lock onto where it needs to aim with more precision.
What this does is use the human eye's innate ability to track objects to solve the target tracking issues, allowing our engineers to simply look at a target and press a button to activate the device.
Importantly, this system could be utilized in vehicles and smart weapons technology, to provide precision aiming without sacrificing the human element.
Logged
Quote from: evicted Saint on discord
Weaponizing Jesus isn't something you do turn 4
Quote from: Alice on a different discord, to iridium, kind of.
hold on, let me keep blowing kisses at him until he stops
My Power armor arms race

Madman198237

  • Bay Watcher
    • View Profile

I am very much not confused. The idea of identifying weak points requires the system to either be able to identify structures that correspond to weak points *or* identify the target and thus be able to point at a specific part of the target.

The system, as described, absolutely requires on picking out enough features and making an identification of what the targeted object is. Your system would certainly be capable of picking out a speed and position, though using a terrible method that should really just be a laser rangefinder (or two, to give a parallax measurement of distance).

To make something even simpler (and probably less controversial for whatever), I suppose we could also just have the engineer mark a point for just long enough for the system to begin tracking the movement of the target using the rangefinders, then it just corrects for the movement the rangefinders see as it fires...
Logged
We shall make the highest quality of quality quantities of soldiers with quantities of quality.

dgr11897

  • Bay Watcher
    • View Profile

I am very much not confused. The idea of identifying weak points requires the system to either be able to identify structures that correspond to weak points *or* identify the target and thus be able to point at a specific part of the target.

The system, as described, absolutely requires on picking out enough features and making an identification of what the targeted object is. Your system would certainly be capable of picking out a speed and position, though using a terrible method that should really just be a laser rangefinder (or two, to give a parallax measurement of distance).

To make something even simpler (and probably less controversial for whatever), I suppose we could also just have the engineer mark a point for just long enough for the system to begin tracking the movement of the target using the rangefinders, then it just corrects for the movement the rangefinders see as it fires...
that would work,
Logged
Quote from: evicted Saint on discord
Weaponizing Jesus isn't something you do turn 4
Quote from: Alice on a different discord, to iridium, kind of.
hold on, let me keep blowing kisses at him until he stops
My Power armor arms race

Madman198237

  • Bay Watcher
    • View Profile

I edited my previous proposal to use the laser rangefinders instead of the x-marks-the-spot option. It makes more sense and is very simple.
Logged
We shall make the highest quality of quality quantities of soldiers with quantities of quality.

Jerick

  • Bay Watcher
    • View Profile

Quote from: Votebox
Heavy Assault Gear
Scourger revision project (1): frostgiant
Basic Overshield Generator - Secondary Power Core (3): frostgiant, Happerry, DGR
Bane Heavy Autolaser (1): Madman

Light Assault Gear
Basic Jump Pack - Variable Nozzle Array (3): Happerry, Jerick, DGR

Engineer Gear
Illuminator Target Illumination Device (2): Madman, Jerick
Autotarget Illuminator Autobrain(0):
Well that addresses my complaints. Switching vote.
Logged

dgr11897

  • Bay Watcher
    • View Profile

Quote from: Votebox
Heavy Assault Gear
Scourger revision project (1): frostgiant
Basic Overshield Generator - Secondary Power Core (3): frostgiant, Happerry, DGR
Bane Heavy Autolaser (1): Madman

Light Assault Gear
Basic Jump Pack - Variable Nozzle Array (2): Happerry, Jerick,

Engineer Gear
Illuminator Target Illumination Device (3): Madman, Jerick, DGR
Autotarget Illuminator Autobrain(0):
« Last Edit: May 22, 2019, 06:05:12 pm by dgr11897 »
Logged
Quote from: evicted Saint on discord
Weaponizing Jesus isn't something you do turn 4
Quote from: Alice on a different discord, to iridium, kind of.
hold on, let me keep blowing kisses at him until he stops
My Power armor arms race

Man of Paper

  • Bay Watcher
    • View Profile
WarBiz Arms Race // Victory Services Thread (Prewar 2 Design Phase)
« Reply #57 on: May 22, 2019, 10:00:13 pm »

Prewar 1 Revision Phase


Proposal: Basic Overshield Generator - Secondary Power Core
Difficulty: Average
Result: (4+2)+0=6, Average

We've taken our Basic Overshield Generator and given it a solid increase in staying power. While doubling the capacity of the Generator does increase size a good bit, it gives Heavy Assaults a second barrier that can be deployed at any time. Due to the additional size of the module the secondary power core is an optional upgrade. Notably however, the power of both shields does increase while both cores are active. While hits will drain both cores faster, the damage they can withstand before falling is greater - during tests a "dummy" often survived a direct high explosive blast without a scratch.

The Basic Overshield Generator - Secondary Power Core is an Auxiliary Item only usable by Heavy Assaults and is considered (EXPENSIVE).


Proposal: Illuminator Target Illumination Device
Difficulty: Average
Result: (6+3)+0=9, Above Average

The Illuminator Target Illumination Device, as redundant as the name sounds, is quite the useful piece of equipment. When an Engineer marks a target they only need to maintain it briefly before the turret begins tracking on it's own. With the turret locked onto it's own stable, consistent marks thanks to it's own rangefinding and targeting abilities it is more able to maintain enough precision to allow it's beam to more readily melt through armor.

This upgrade to the IAALT sees a greater acceptance of the utility of the device, reflected in it's expense level dropping to (EXPENSIVE) as willingness to field it goes up while overall cost is barely impacted.


----------------


INCOMING ALERT

You are in your second DESIGN PHASE. Once again you have two designs to vote on, however some noted the absence of the MAX class available in PS2. This was intentional! Instead of forcing Exoskeletons on everyone, you are being asked instead to use one design to create a special class for your team. This means both defining the role of the class on the battlefield, as well as creation of it's basic class-defining equipment! I will aid teams in figuring out what they can work with, to a point. The special Class design will be slightly easier than normal and will also be granted a reroll as if it had received a Research Credit.


Spoiler: VS Armory (click to show/hide)
Logged

Jerick

  • Bay Watcher
    • View Profile
Re: WarBiz Arms Race // Victory Services Thread (Prewar 2 Design Phase)
« Reply #58 on: May 23, 2019, 09:11:18 am »

Quote from: Operator Special Class
The operator is a specialist class focused on information gathering and warfare. Aside from signal interception gear they also heavily utilise remote or autonomous systems which they connect to and control through their powerful computer and communications equipment. They role on the battlefield is not to directly engage the enemy but to deploy drones to survey the enemy and to use their comn systems and computers to intercept and decrypt enemy comns and to provide flexible real time info for nearby friendlies.

Gear
Siren Comns unit:
The siren comns unit is a comprehensive package of computers, comunication receivers, sensors for picking up enemy transmissions and communications jamming gear. The Siren has also the capability to command and remotely control drones. The long range of the siren gives any connected drone impressive range from the user. The Siren's computers are very powerful and provide an excellent platform for decryption of staging hacking attempts at enemy equipment.

Sky Eye Light Recon drone: The Sky Eye is a small disposable flying drone. It is very small and has four rotors keeping it stable and flying. It is remotely controlled by the operator who receives a video feed from the drone's high resolution cameras. It has a small built in explosive charge that can be triggered by the operator at any time in order to both destroy the drone and deal damage to anything the drone is right next to. However the small size of the drone and thus the charge limits it's effectiveness. These drones can be manufactured on site with nanites similar to how engineers build turrets, with a cooldown time before the operator is able to construct another one. However the small size of the Sky Eye means the cooldown time of making another is quite short. Larger and more complicated drones will likely take longer.
My suggestion for our special class.
Logged

Happerry

  • Bay Watcher
    • View Profile
Re: WarBiz Arms Race // Victory Services Thread (Prewar 1 Design Phase)
« Reply #59 on: May 25, 2019, 02:38:33 am »

Since everything has been silent for a bit, I'm putting this votebox up, with at least the equipment designs even if we still can't agree on a good class to design. Voting for the Electrocarbine because we have 0 weapons that cover the close range, and we really should have some for close ranged fights in general and for the Light Assault in specific, and for the Storm Hammer because we need a non-turret based anti-vehicle weapon for our Heavy Assaults to use.

Quote from: Votebox.
Infiltrator Gear
Hellstrike Plasma Knife (0):

Heavy Assault Gear
Storm Hammer Energy Torpedo Projector (1): Happerry

Light Assault Gear
Spark Electrocarbine (1): Happerry

Combat Medic Gear
Bubble Barrier Bombs (0):

Engineer Gear
Muninn Surveillance UAVs (0):
Pylon Point Defense Tower (0):

All Class Gear
Blue Burst Magnet Grenade (0):
Pulser Laser Pistol (0):
Sunspitter Pistol (0):
Light Charger Laser Grenade/Mine (0):
Nameless Death SMG (0):
Logged
Forenia Forever!
GENERATION 11: The first time you see this, copy it into your sig on any forum and add 1 to the generation. Social experiment.
Pages: 1 2 3 [4] 5 6 ... 18