Web ui mockup

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ ┃ 🐕 SHADOWHOUND DASHBOARD ┃ ┃ AUTONOMOUS ROBOT MISSION CONTROL ┃ ┃ [CONNECTED] ┃ ┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

┌─────────────────────────────────────┬─────────────────────────────────────┐ │ 📹 CAMERA FEED │ ⚡ PERFORMANCE METRICS │ │ │ │ │ ┌───────────────────────────────┐ │ ┌───────────┬────────────────────┐│ │ │ │ │ │ AGENT │ 1.23s 🟢 ││ │ │ [Camera Feed Image] │ │ ├───────────┼────────────────────┤│ │ │ │ │ │ OVERHEAD │ 0.045s 🟢 ││ │ │ │ │ ├───────────┼────────────────────┤│ │ │ │ │ │ TOTAL │ 1.28s 🟢 ││ │ │ │ │ ├───────────┼────────────────────┤│ │ │ │ │ │ COMMANDS │ 42 ││ │ └───────────────────────────────┘ │ └───────────┴────────────────────┘│ │ │ │ │ │ Averages (Last 50 Commands) │ │ │ ┌───────────┬────────────────────┐│ │ │ │ AVG AGENT │ 1.45s ││ │ │ ├───────────┼────────────────────┤│ │ │ │ AVG TOTAL │ 1.52s ││ │ │ └───────────┴────────────────────┘│ └─────────────────────────────────────┴─────────────────────────────────────┘

┌─────────────────────────────────────┬─────────────────────────────────────┐ │ 📊 DIAGNOSTICS │ 💻 TERMINAL │ │ │ │ │ ROBOT MODE: STANDING │ [14:32:10] > take one step forward│ │ TOPICS: 12 ACTIONS: 3 │ [14:32:12] ✅ Robot moved forward │ │ LAST UPDATE: 14:32:15 │ [14:32:12] ⏱️ TIMING: Agent │ │ │ 1.23s | Overhead │ │ Available Topics: │ 0.045s | Total 1.28s │ │ /cmd_vel /odom /camera/image_raw │ [14:32:15] > rotate 90 degrees │ │ /nav2/goal /status │ [14:32:17] ✅ Robot rotated 90° │ │ │ [14:32:17] ⏱️ TIMING: Agent │ │ │ 1.15s | Overhead │ │ │ 0.038s | Total 1.19s │ │ │ [14:32:20] > stand up │ └─────────────────────────────────────┴─────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────────────────────┐ │ 📡 MISSION STATUS │ │ ┌───────────────────────────────────────────────────────────────────────┐ │ │ │ ✅ Robot rotated 90 degrees successfully │ │ │ └───────────────────────────────────────────────────────────────────────┘ │ └─────────────────────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────────────────────┐ │ 🎮 COMMAND CENTER │ │ │ │ Custom Command │ │ ┌───────────────────────────────────────────────────────────────────────┐ │ │ │ Enter mission command (e.g., 'stand up and wave hello') │ │ │ └───────────────────────────────────────────────────────────────────────┘ │ │ [▶️ EXECUTE] │ │ │ │ Quick Commands │ │ [🧍 STAND] [🪑 SIT] [👋 WAVE] [💃 DANCE] [🤸 STRETCH] [⚖️ BALANCE] │ └─────────────────────────────────────────────────────────────────────────────┘

═══════════════════════════════════════════════════════════════════════════════ PERFORMANCE INSIGHTS ═══════════════════════════════════════════════════════════════════════════════

COLOR CODING: 🟢 Green = Good (Agent <2s, Total <3s) 🟡 Yellow = Warning (Agent 2-4s, Total 3-5s) ← Consider optimization 🔴 Red = Slow (Agent >4s, Total >5s) ← Needs optimization!

TIMING BREAKDOWN: Agent Duration = Time in DIMOS agent (LLM + skill execution) Overhead Duration = ROS messaging + Python execution time Total Duration = End-to-end mission execution time

EXAMPLE OUTPUTS IN TERMINAL:

After sending "take one step forward": ┌────────────────────────────────────────────────────────────────────┐ │ [14:32:10] > take one step forward │ │ [14:32:12] ✅ Robot moved forward successfully │ │ [14:32:12] ⏱️ TIMING: Agent 1.23s | Overhead 0.045s | Total 1.28s│ └────────────────────────────────────────────────────────────────────┘

After sending "go to the kitchen": ┌────────────────────────────────────────────────────────────────────┐ │ [14:35:20] > go to the kitchen │ │ [14:35:24] ✅ Navigating to kitchen waypoint │ │ [14:35:24] ⏱️ TIMING: Agent 3.45s | Overhead 0.089s | Total 3.54s│ └────────────────────────────────────────────────────────────────────┘

After sending "describe what you see" (with VLM): ┌────────────────────────────────────────────────────────────────────┐ │ [14:40:10] > describe what you see │ │ [14:40:15] ✅ I see a hallway with white walls... │ │ [14:40:15] ⏱️ TIMING: Agent 4.82s | Overhead 0.103s | Total 4.92s│ └────────────────────────────────────────────────────────────────────┘ ↑ VLM queries add 1-3s to agent time!

═══════════════════════════════════════════════════════════════════════════════ WHAT YOU'LL SEE ═══════════════════════════════════════════════════════════════════════════════

  1. IMMEDIATE FEEDBACK in terminal:
  2. Command echoed with timestamp
  3. Success/failure message

  4. Timing breakdown (⏱️ emoji makes it easy to spot)

  5. REAL-TIME METRICS in performance panel:

  6. Latest timing values update every second
  7. Color-coded indicators for quick assessment

  8. Running averages smooth out outliers

  9. TREND ANALYSIS:

  10. Watch averages stabilize over multiple commands
  11. Identify if performance degrades over time

  12. Compare simple vs complex commands

  13. BOTTLENECK IDENTIFICATION:

  14. If Agent ≈ Total → Cloud API bottleneck
  15. If Overhead > 10% → ROS/Python bottleneck
  16. If Agent varies wildly → DIMOS skill bottleneck

═══════════════════════════════════════════════════════════════════════════════ NEXT STEPS ═══════════════════════════════════════════════════════════════════════════════

  1. Start the system: ros2 launch shadowhound_bringup shadowhound_bringup.launch.py

  2. Open browser to http://localhost:8080

  3. Send test commands and watch metrics update

  4. Collect baseline data:

  5. Simple commands: "stand", "sit", "wave"
  6. Navigation: "go forward", "rotate 90"

  7. Complex: "patrol hallway"

  8. Analyze results:

  9. What's the average agent duration?
  10. Is cloud API the bottleneck? (likely yes)

  11. Should we optimize before adding VLM? (yes!)

  12. Optimize based on findings:

  13. If agent >2s → Try gpt-3.5-turbo (faster, cheaper)
  14. If overhead high → Profile ROS callbacks

  15. Add progress indicators for better UX

  16. Then integrate VLM with optimized baseline! 🚀