DrAmaT - Editor - AI Agent examples

« AI Agent window (editor) | Contents | AI Agent window of the Robot Cell Simulator »

This page shows examples for the AI Agent window of the editor: real prompts and what the agent did with them, complete checks with the full text of the answer and the report of a VKRC robot check of five robots. The functions of the window, its commands, the knowledge files and the settings are described on the page AI Agent window (editor). The agent answers in the language of the prompt.

Prompt examples Examples: one prompt – a complete check Example 1 – points Example 2 – spot welds Example 3 – VKRC robot check of 5 robots Example 4 – VFanuc robot check of 4 robots Example 5 – writing Makro45

DReS Editor: Teach Your AI Agent - Your Own Knowledge Files for KUKA VKRC Robot Checks (YouTube, English) – with Example 5 (Makro45)
Deutsch

DReS Editor: Den KI-Agenten anlernen - eigene Wissensdateien für die Roboterprüfung von KUKA VKRC (YouTube, Deutsch)

Prompt examples

Real prompts and what the agent did with them – one sentence in plain words, the agent finds the way and the commands itself. All results below come from real runs; the names of robots, stations and plants are replaced (R01, R02 … or the example station 0001R01 … 0001R05). Click a picture to see it in full size.

Collision map of a whole station from robot archives

PromptGenerate collision map for all robots archives in D:\TEST. Check any logic problems.

What the agent did:

opened all robot archives of the folder in the editor, one after another (16 Fanuc robots);
read every Folge and UP: the requests DO[80+n] with CALL MAKRO020, the release inputs DI[40+n] in TC_ONLINE, the releases DO[40+n] and the markers of MAKRO050;
found the partner of every zone in the comments and long texts and checked that both robots of a pair use the same zone number;
drew the map: robots, lines with the zone numbers – green consistent, orange wrong comment / partner, red error in the code, grey partner outside the archives – with a panel listing every finding with program and line.

Result: the interlock map of 16 Fanuc robots – no program was changed, the agent only reads and reports. Examples of the findings: a zone requested and never released, a release input missing in TC_ONLINE, comments naming the wrong partner robot, zones held to the end of the UP.

16Fanuc robots

30robot pairs consistent

5errors in the code

7wrong comments / partners

The map drawn by the agent: 16 Fanuc robots on two lines, zone numbers on the lines, findings at the right (robot names replaced by R01…R17)

Now also a function of the editor: Tools > Collision map of the robot projects and the API command collision_map build the same map for KUKA VKRC and Fanuc stations in one step – picture, HTML / PDF report and CSV / Excel tables. It is a signal analysis of the programs; the geometry is checked in the Robot Cell Simulator (work zones, Live 3D).

Points of the Folgen and of the UP calls

PromptCheck whether points 1, 2 and the last one in every Folge have the same coordinates, and whether the first point of every UP has the same coordinates as the point where the UP is called. The difference may be max. 2 mm. Show me the programs where this condition is not met.

Result: 38 Folgen of 5 robots and 118 UP calls in about 1.5 minutes. Two calls are 5.0 mm off – compared in world coordinates, also with different bases; everything else OK. The agent names the likely cause (UP42 itself or the definition of Base 1) and changes nothing. Full answer and picture: Example 1.

Spot welds: SERVOZANGE parameters

PromptFind the SERVOZANGE commands and show me all used Programmnummer, Punktnummer and FahrzeugTyp. Check if Offset is greater than 0, only in UP programs. Mark the values 0 in red.

Result: 143 SERVOZANGE commands of all robot projects in one table, read from the disk – no file opened, nothing changed. Offset 30 in one robot, Offset 0 in two robots (50 welds, in red); the agent also noticed by itself weld points used on two robots. Full answer and picture: Example 2.

Robot Cell Simulator: cycle times and waits

Prompt (Robot Cell Simulator)Run Folge 81 once on all robots and give me the cycle times: which robot sets the cycle and what do the robots wait for?

Result: the agent ran Folge 81 once on all 5 robots (400 %, SPS off): station cycle time 69.1 s, set by R03; table of the cycle times and the interlock waits of every robot, the waits with program and point, a request race on zone 1 (the simulator keeps the order as an automatic race rule), contacts with and without an interlock (also in 3D) and the programs that were not simulated – and it proposes the next step (shared work areas and a robustness test). Details: AI Agent window of the Robot Cell Simulator.

The answer of the agent in the Robot Cell Simulator (anonymized example station 0001R01…0001R05)

Robot Cell Simulator: all possible collisions of a station

Prompt (Robot Cell Simulator)Find all possible collisions of the open station: baseline run, work zones, check interlocks, then write and run a robustness test plan with Live 3D. Give me the report.

What the agent did: all robots to HOME, baseline run, work zones, Check interlocks, a test plan written from the results of the baseline run, the robustness test with Live 3D and the report – one prompt instead of eight manual steps. Result on a real station (anonymized, 5 robots, known collision places A–E):

Test	Time	Collision places found
fixed plan (engineer)	18 min	A, B, D
random generator	45 min	B, E
plan of the AI agent	10 min	A, B, C, D, E

The agent also found request races by itself: the order of two robots at one zone decides about 10 s of the station cycle time (72 s → 82 s) – it tests the other order on purpose. Details: AI Agent window of the Robot Cell Simulator.

Examples: one prompt – a complete check

The two examples below were made on an anonymized example station with five robot projects (0001R01 to 0001R05). Each shows the prompt as it was typed, the editor with the answer and the full text of the answer. The answer text contains the markers {red}…{/red} and {green}…{/green}: the chat shows these parts in color (see Colors); in the text below they are left as they are.

Example 1 – points of the Folgen and of the UP calls

Prompt:

Check whether points 1, 2 and the last one in every Folge have the same coordinates, and whether the first point of every UP has the same coordinates as the point where the UP is called. The difference may be max. 2 mm. Show me the programs where this condition is not met.

The agent says that this takes a moment and runs one points call over the Folgen of all robot projects (fileMask ["folge*"]) with the checks samePoints ([["1","2"],["1","last"]]) and callStart, tolerance 2 mm. DReS computes the distances itself; points with different bases and the same tool are compared in world coordinates (bases from $config.dat), points with different tools are reported as not comparable. It takes about 1.5 minutes (about 2 s per program).

Result: 38 Folgen of 5 robots – 76 point checks, 0 failed; 118 UP calls, 2 failed: in 0001R05 FOLGE81 and Folge83 call UP42 at P2, and the first point of UP42 differs by 5.0 mm (compared in world coordinates, Base 0 against Base 1). Because both Folgen show the same difference, the agent concludes that UP42 itself (its P1 or the definition of Base 1) is the likely cause. Nothing was changed. (In this example station one point was moved by 5 mm on purpose to show how a difference is found.)

The editor with the AI Agent window: result of example 1 – 5.0 mm in red, 0 failed point checks in green

Full answer (as in the chat):

You: Check whether points 1, 2 and the last one in every Folge have the same coordinates, and whether the first point of every UP has the same coordinates as the point where the UP is called. The difference may be max. 2 mm. Show me the programs where this condition is not met.

Agent: This takes a moment, roughly 2 s per program.

Agent: Two programs fail the check. Both fail at the same spot: the call of UP42 (the first point of a UP vs the point where it is called).

I checked 38 Folgen across all robot projects with a 2 mm tolerance.

| Check | Checked | Failed |
|---|---|---|
| Points 1 = 2 and 1 = last point in each Folge | 76 | {green}0{/green} |
| First point of the UP = the point where it is called | 118 calls | {red}2{/red} |

**Programs that fail:**

| Project | Program | Line | Call | Point of the call | First point of the UP | Difference |
|---|---|---|---|---|---|---|
| 0001R05 | KRC/R1/Folgen/FOLGE81.src | 39 | UP42 | P2 | P1 | {red}5.0 mm{/red} |
| 0001R05 | KRC/R1/Folgen/Folge83.src | 39 | UP42 | P2 | P1 | {red}5.0 mm{/red} |

- In both Folgen, P2 uses Tool 1 / Base 0 and the first point of UP42 uses Tool 1 / Base 1. They were compared in world coordinates.
- Both programs show the same 5 mm difference, so UP42 itself is the likely cause: its first point P1, or the definition of Base 1. It's less likely that both Folgen have the same wrong P2.
- In every Folge, points 1, 2 and the last one have the same coordinates.

I didn't change anything.

Example 2 – spot welds (SERVOZANGE)

Prompt:

Find the SERVOZANGE commands and show me all used Programmnummer, Punktnummer and FahrzeugTyp. Check if Offset is greater than 0, only in UP programs. Mark the values 0 in red.

The agent runs tech list with scope projects, fields (Programmnummer, Punktnummer, FahrzeugTyp, Offset) and fileMask ["up*"]. It reads the files from the disk – nothing is opened, nothing is changed – and gets a summary of the used values with their counts, so even hundreds of spot welds fit into a short answer.

Result: 143 SERVOZANGE commands in the UP programs of 0001R01, 0001R04 and 0001R05, of them 98 Schweissen: Programmnummer 99 everywhere, FahrzeugTyp 6 or 7; Offset 30 only in 0001R01 (48 welds), Offset 0 (red) in 0001R04 and 0001R05 (50 welds). The service commands Setzhub, Fraesen and Service have FahrzeugTyp 0. The agent also noticed by itself that some Punktnummer are used in two robots and asks to check them. At the end it offers a CSV export or a change of the Offset values (with a preview first).

Result of example 2: the table of the spot welds – Offset 0 in red, yes green / no red

Full answer (as in the chat):

You: Find the SERVOZANGE commands and show me all used Programmnummer, Punktnummer and FahrzeugTyp. Check if Offset is greater than 0, only in UP programs. Mark the values 0 in red.

Agent: I found 143 SERVOZANGE commands in the UP programs of 0001R01, 0001R04 and 0001R05. Of these, 98 are Schweissen (spot welding) commands. Only those have a Programmnummer and an Offset. I only read the files from disk, opened nothing and changed nothing.

**Schweissen commands (Offset check)**

| Project | UP | Programmnummer | FahrzeugTyp | Punktnummer | Count | Offset | Offset > 0 |
|---|---|---|---|---|---|---|---|
| 0001R01 | UP42 | 99 | 6 | 15068–15073 | 4 | 30 | {green}yes{/green} |
| 0001R01 | Up43 | 99 | 6 | 15920–16019 | 10 | 30 | {green}yes{/green} |
| 0001R01 | UP44 | 99 | 6 | 15874–15999 | 10 | 30 | {green}yes{/green} |
| 0001R01 | UP62 | 99 | 7 | 15066–15071 | 4 | 30 | {green}yes{/green} |
| 0001R01 | UP63 | 99 | 7 | 15860–16037 | 10 | 30 | {green}yes{/green} |
| 0001R01 | UP64 | 99 | 7 | 15814–15953 | 10 | 30 | {green}yes{/green} |
| 0001R04 | Up42 | 99 | 6 | 15324–15329 | 6 | {red}0{/red} | {red}no{/red} |
| 0001R05 | UP42 | 99 | 6 | 15846–15872 | 22 | {red}0{/red} | {red}no{/red} |
| 0001R05 | UP62 | 99 | 7 | 15786–15812 | 22 | {red}0{/red} | {red}no{/red} |

**Other SERVOZANGE commands** (UP101, UP112, UP201 and UP221 in all three projects; these commands have no Programmnummer)

| Command | Punktnummer | FahrzeugTyp | Count |
|---|---|---|---|
| Setzhub | 61 / 71 | {red}0{/red} | 18 |
| Fraesen | 31 | {red}0{/red} | 3 |
| Service | i59 | {red}0{/red} | 3 |

The Messen, Andocken and Abdocken commands have no parameters.

**Results**
- **Offset:** All 48 welds in 0001R01 have Offset 30. All 50 welds in 0001R04 and 0001R05 have {red}Offset 0{/red}.
- **Values used:** Every weld uses Programmnummer 99. FahrzeugTyp is 6 (52 welds) or 7 (46 welds).
- **Repeated weld points:** Points 15860, 15861 and 15864–15867 are used twice each, and so are 15870 and 15871. They appear in 0001R01 UP63 and again in 0001R05 UP42. Please check whether these weld points really should be on two robots.

I can export the full list to a CSV file for Excel, or change the Offset values (I'd show you a preview first).

Further examples of messages:

Show the robot projects and set R03 as active
Open folge81 of R03 at line 40 and check its syntax
Show the VW user, software version and IP of all robots
Which tools and bases are used in the Folgen and are they all defined in the robot data?
Set FahrzeugTyp to 2 in all spot welds of all robot projects – first show me what would change
Save all modified files and start the Robot Cell Simulator

Example 3 – VKRC robot check of 5 robots

The agent checks the robot data of five VKRC robot projects of one session like the plugin VKRC Robot Check and more: called but missing Makros and UPs, tools, bases and loads, calibration and the software limit switches including the 7th axis – and writes a report with a comparison of the robots. It only reads the robot data; it does not change or delete files.

Prompt:

Do a VKRC robot check of all robot projects in this session: Makros, UPs, tools, bases, loads, calibration and software limit switches (also the 7th axis), and write a short report with what the programmer must correct, with a comparison of the robots.

The report of the agent – scroll in the frame or open the report (PDF written by the agent). On request the agent saves it as a PDF, HTML or Excel file (e.g. Save it as PDF on the desktop).

Tokens: a check of many robots reads hundreds of programs and costs many tokens of your AI account (5 robots with about 340 Makros: over 2 million input tokens). For a quick check give one or two robots per chat and start a new chat for the next ones.

More about the check and other prompts: VKRC Robot Check > Robot check with the AI agent.

Example 4 – VFanuc robot check of 4 robots

The same check for Fanuc backups in the VW standard: Folgen, UPs and Makros (called but missing, never called, programs only as .TP without .LS), tools UT and user frames UF used in the points, payload schedules (PAYLOAD[n]) against the rated payload of the robot, mastering (SYSMAST.VA), the software limit switches and the travel track E1 with the points outside its limits, points not taught. The agent reads VERSION.VA, SYMOTN.VA, SYSFRAME.VA, SYSMAST.VA and the LS programs; a missing file is named in the report.

Prompt:

Do a VFanuc robot check of the active Fanuc project: Makros, UPs, tools UT, user frames UF, payloads, mastering and software limit switches (also the travel track E1), and write a short report with what the programmer must correct.

For several robots: ... of all Fanuc robot projects in this session ... with a comparison of the robots.

The report of the agent for 4 Fanuc robots of one session – prompt:

Prompt:

Do a VFanuc robot check of all Fanuc robot projects in this session: Makros, UPs, tools UT, user frames UF, payloads, mastering and software limit switches (also the travel track E1), and write a short report with what the programmer must correct, with a comparison of the robots.

Scroll in the frame or open the report (PDF written by the agent).

Example 5 – writing a program: Makro45 with the collision releases

The agent does not only check programs – it also writes them. In this example it creates the Makro45 of a KUKA VKRC robot with the releases of exactly those collision areas that the robot requests in its programs, and writes the long text of every signal as a comment above the command.

Prompt:

Create Makro45 for robot KDA4G1214460R01 with only the signals of the collision areas the robot uses in its programs (A41-A56 for its requests A81-A96) and set them to EIN; above every command add a comment with the long text of the variable.

What the agent did:

searched the programs of the robot for the requests A81–A96 and found only A81–A85 (23 requests in UP42, UP43, UP44, UP62, UP63, UP64 and UP201) – so the Makro sets only the matching releases A41–A45, A46–A56 are not in it;
took the header and the format from the existing Makros of the robot (Makro40, Makro61; TPVW 8.7.2, DEFFCT INT MAKRO45(ADV :IN));
read the long texts of A41–A45 from the robot data (RefListe.txt) and wrote Makro45 with A41 = EIN … A45 = EIN and a comment with the long text above every output;
opened the Makro in the editor and checked its syntax (check_syntax – no errors or warnings); like the other Makros of this robot it has only a .src file;
told what to check: the header text written by the agent and that no Folge or UP calls Makro45 yet.

Makro45 written by the agent: one comment with the long text above every release A41–A45

The answer of the agent in the AI Agent window (the line with the path of the file is blurred)

Prompts work in any language: write the prompt in your language – the agent answers in the same language. The same task in German: Beispiel 5.

« AI Agent window (editor) | Contents | AI Agent window of the Robot Cell Simulator »