Chapter 6: Phase 3 - Communication Protocol

6.1 Phase 3 Overview and Advanced Capabilities

Phase 3 introduces MQTT-based messaging protocols enabling sophisticated real-time coordination between robots, edge computing infrastructure, and central management systems. This phase unlocks advanced scenarios like swarm cleaning, dynamic task reallocation, and 5G-enabled fleet orchestration.

Beyond Request-Response

While Phase 2 APIs excel at direct robot control, they follow request-response patterns unsuitable for real-time multi-robot coordination. Phase 3 adopts publish-subscribe messaging where robots, controllers, and edge systems communicate asynchronously through message brokers.

Key Capabilities Enabled

• Swarm Coordination: Multiple robots negotiate task distribution and avoid conflicts
• Dynamic Reallocation: Tasks reassigned instantly when robots encounter problems
• Edge Intelligence: AI models deployed and synchronized across robot fleets
• 5G Integration: Ultra-low latency coordination through network slicing
• Predictive Operations: Anticipate maintenance needs and optimize schedules

6.2 MQTT Protocol Foundation

MQTT (Message Queuing Telemetry Transport) serves as the Phase 3 protocol foundation. Originally designed for IoT devices with limited bandwidth and unreliable networks, MQTT provides lightweight publish-subscribe messaging ideal for cleaning robot coordination.

MQTT Advantages for Robotics

Feature	Benefit	Robotics Application
Publish-Subscribe	Decoupled communication	Robots broadcast status, subscribers receive updates
Quality of Service	Guaranteed delivery levels	Critical commands delivered exactly once
Persistent Sessions	Survive disconnections	Robots reconnect without losing state
Last Will Testament	Automatic disconnect detection	Fleet notified when robot goes offline
Retained Messages	Late subscribers get state	New controllers receive current robot status
Lightweight Protocol	Minimal bandwidth usage	Works over cellular, WiFi, Ethernet equally

QoS Levels for Robot Messaging

6.3 Topic Namespace Design

MQTT topics organize messages hierarchically. WIA-ROB-011 defines a standardized namespace enabling logical message routing and access control.

Topic Hierarchy

wia/rob-011/{version}/{facility}/{area}/{robot_id}/{category}/{subcategory}

Examples:
wia/rob-011/v1/building-a/floor-1/robot-001/status/battery
wia/rob-011/v1/building-a/floor-1/robot-001/telemetry/position
wia/rob-011/v1/building-a/floor-1/robot-001/commands/start
wia/rob-011/v1/building-a/floor-1/+/status/online
wia/rob-011/v1/building-a/+/+/alerts/error
wia/rob-011/v1/+/+/+/fleet/coordination

Standard Topic Categories

Category	Direction	QoS	Purpose
status/*	Robot → Subscribers	1	State changes (online, cleaning, error)
telemetry/*	Robot → Subscribers	0	High-frequency sensor data
commands/*	Controller → Robot	2	Control instructions
alerts/*	Robot → Subscribers	1	Warnings, errors, maintenance needs
fleet/*	Bidirectional	2	Multi-robot coordination
edge/*	Edge → Robot	1	AI model updates, parameters

6.4 Real-Time Navigation Coordination

When multiple robots operate in shared spaces, collision avoidance and path optimization require continuous position sharing and negotiation.

Position Broadcasting

Topic: wia/rob-011/v1/warehouse/zone-a/robot-001/telemetry/position
QoS: 0 (high frequency acceptable loss)
Frequency: 5 Hz

Payload:
{
  "timestamp": "2025-01-15T14:30:15.234Z",
  "position": {
    "x": 15.23,
    "y": 8.47,
    "z": 0.0,
    "theta": 1.571
  },
  "velocity": {
    "linear": 0.35,
    "angular": 0.0
  },
  "trajectory": {
    "waypoints": [
      {"x": 15.5, "y": 8.5, "eta_ms": 500},
      {"x": 16.0, "y": 8.5, "eta_ms": 1500}
    ]
  },
  "state": "navigating"
}

Collision Avoidance Protocol

Robots subscribe to position topics of nearby robots. When trajectories intersect, robots negotiate right-of-way based on priority rules:

Topic: wia/rob-011/v1/warehouse/zone-a/fleet/collision-avoidance
QoS: 2

Robot-001 detects potential collision with Robot-002:
{
  "type": "collision_warning",
  "robot_id": "robot-001",
  "conflicting_robot": "robot-002",
  "intersection_point": {"x": 16.0, "y": 8.5},
  "time_to_collision_ms": 2000,
  "priority": 5,
  "proposed_action": "slow_and_yield"
}

Robot-002 responds:
{
  "type": "collision_response",
  "robot_id": "robot-002",
  "priority": 3,
  "action": "maintain_course"
}

Robot-001 adjusts:
{
  "type": "collision_resolved",
  "robot_id": "robot-001",
  "action_taken": "stopped_at_safe_distance"
}

6.5 Fleet Management and Task Coordination

Commercial cleaning operations require intelligent task distribution, load balancing, and dynamic reallocation when robots encounter problems or complete work faster than expected.

Task Assignment Protocol

Topic: wia/rob-011/v1/building-a/floor-1/fleet/tasks
QoS: 2

Fleet coordinator publishes task:
{
  "task_id": "task-uuid-789",
  "type": "cleaning_task",
  "area_id": "conference-room-b",
  "priority": "high",
  "deadline": "2025-01-15T17:00:00Z",
  "requirements": {
    "mode": "vacuum_and_mop",
    "surface_types": ["carpet", "tile"],
    "estimated_duration_minutes": 25
  },
  "constraints": {
    "requires_full_battery": true,
    "avoid_occupied_spaces": true
  },
  "status": "available"
}

Robots evaluate and bid:
{
  "task_id": "task-uuid-789",
  "robot_id": "robot-003",
  "bid": {
    "can_complete": true,
    "estimated_start": "2025-01-15T16:30:00Z",
    "estimated_completion": "2025-01-15T16:55:00Z",
    "battery_at_start": 85,
    "current_distance_meters": 45,
    "priority_score": 0.87
  }
}

Coordinator assigns task:
{
  "task_id": "task-uuid-789",
  "assigned_to": "robot-003",
  "status": "assigned",
  "confirmed_start": "2025-01-15T16:30:00Z"
}

Dynamic Reallocation

When robots encounter obstacles, battery depletion, or mechanical issues, tasks automatically reassign to available robots:

Robot-003 encounters problem:
{
  "task_id": "task-uuid-789",
  "robot_id": "robot-003",
  "status": "unable_to_complete",
  "reason": "obstacle_blocking_area",
  "completion_percent": 30,
  "work_remaining": {
    "area_id": "conference-room-b",
    "sections_remaining": ["section-2", "section-3"]
  }
}

Coordinator reassigns:
{
  "task_id": "task-uuid-789-continuation",
  "original_task": "task-uuid-789",
  "assigned_to": "robot-005",
  "work_scope": "sections_remaining",
  "priority": "urgent"
}

6.6 Edge Computing Integration

Edge servers deployed in facilities provide computational resources for AI inference, data aggregation, and fleet optimization. Phase 3 protocol integrates robots with edge infrastructure.

AI Model Distribution

Topic: wia/rob-011/v1/building-a/+/edge/model-update
QoS: 1

Edge server publishes updated model:
{
  "model_id": "dirt-detection-v3.2",
  "model_type": "tensorflow_lite",
  "version": "3.2.0",
  "download_url": "https://edge.example.com/models/dirt-v3.2.tflite",
  "checksum_sha256": "abc123...",
  "size_bytes": 8388608,
  "deployment": {
    "target_robots": ["robot-001", "robot-002", "robot-003"],
    "rollout_strategy": "gradual",
    "activation_time": "2025-01-16T02:00:00Z"
  },
  "performance_improvements": {
    "accuracy_increase": 0.05,
    "inference_speed_ms": 12
  }
}

Robots acknowledge and download:
{
  "model_id": "dirt-detection-v3.2",
  "robot_id": "robot-001",
  "download_status": "completed",
  "validation_passed": true,
  "ready_for_activation": true
}

Edge-Assisted Path Planning

Complex path optimization calculations offload to edge servers with more computational power than robot hardware:

Robot requests path optimization:
Topic: wia/rob-011/v1/building-a/floor-1/robot-001/edge/path-request

{
  "request_id": "path-req-456",
  "current_position": {"x": 2.0, "y": 3.0},
  "destinations": [
    {"area_id": "room-5", "priority": "high"},
    {"area_id": "room-8", "priority": "medium"},
    {"area_id": "room-3", "priority": "low"}
  ],
  "constraints": {
    "battery_remaining": 65,
    "time_limit_minutes": 45,
    "avoid_areas": ["room-6-occupied"]
  }
}

Edge responds with optimized path:
Topic: wia/rob-011/v1/building-a/floor-1/robot-001/edge/path-response

{
  "request_id": "path-req-456",
  "optimal_path": {
    "waypoints": [...],
    "total_distance_meters": 127.3,
    "estimated_duration_minutes": 42,
    "battery_required": 58,
    "route_efficiency": 0.91
  },
  "alternative_paths": [...]
}

6.7 5G Network Slicing and QoS

5G networks provide guaranteed quality of service through network slicing—dedicated virtual networks with specific latency, bandwidth, and reliability characteristics.

Network Slice Types for Robotics

Slice Type	Latency	Bandwidth	Use Case
Ultra-Reliable Low-Latency (URLLC)	< 10ms	Medium	Emergency stops, collision avoidance
Enhanced Mobile Broadband (eMBB)	20-50ms	High	Video streaming, map updates
Massive IoT (mMTC)	100-500ms	Low	Periodic telemetry, status updates

Network Slice Selection

Robots request appropriate network slice based on operation:

Emergency stop command:
{
  "command": "emergency_stop",
  "network_requirements": {
    "slice_type": "URLLC",
    "max_latency_ms": 10,
    "reliability": 0.999999
  }
}

Routine status update:
{
  "status": "cleaning",
  "network_requirements": {
    "slice_type": "mMTC",
    "max_latency_ms": 500,
    "reliability": 0.99
  }
}

6.8 Swarm Intelligence Algorithms

Multiple robots cleaning collaboratively employ swarm intelligence principles inspired by ant colonies, bee hives, and flocking birds.

Coverage Optimization Through Pheromones

Virtual pheromones mark cleaned areas, preventing redundant coverage while ensuring completeness:

Topic: wia/rob-011/v1/warehouse/zone-a/fleet/pheromones
QoS: 0 (high frequency, ephemeral data)

Robot deposits virtual pheromone after cleaning:
{
  "robot_id": "robot-002",
  "timestamp": "2025-01-15T14:30:45Z",
  "pheromone": {
    "position": {"x": 10.5, "y": 12.3},
    "type": "cleaned",
    "strength": 1.0,
    "decay_rate": 0.1
  }
}

Other robots sense pheromone strength and avoid recently cleaned areas,
naturally distributing across uncleaned space.

Emergent Task Allocation

Rather than centralized task assignment, robots self-organize based on local information:

6.9 Predictive Maintenance Protocol

Robots monitor component health and predict maintenance needs before failures occur, minimizing downtime and extending equipment life.

Health Telemetry

Topic: wia/rob-011/v1/building-a/floor-1/robot-001/health/components
QoS: 1

{
  "timestamp": "2025-01-15T14:30:00Z",
  "components": [
    {
      "component": "main_brush",
      "health_score": 0.65,
      "runtime_hours": 287,
      "expected_lifespan_hours": 300,
      "failure_probability_30days": 0.42,
      "recommended_action": "replace_within_2_weeks"
    },
    {
      "component": "left_wheel_motor",
      "health_score": 0.92,
      "anomalies_detected": [
        {
          "type": "vibration_increase",
          "severity": "low",
          "trend": "worsening"
        }
      ],
      "recommended_action": "monitor"
    }
  ]
}

Maintenance Scheduling Coordination

Fleet coordinator optimizes maintenance timing:
Topic: wia/rob-011/v1/building-a/fleet/maintenance-schedule

{
  "schedule": [
    {
      "robot_id": "robot-001",
      "maintenance_type": "brush_replacement",
      "recommended_window": {
        "start": "2025-01-20T08:00:00Z",
        "end": "2025-01-20T10:00:00Z"
      },
      "urgency": "medium",
      "impact_if_delayed": "reduced_cleaning_efficiency"
    }
  ],
  "optimization_criteria": {
    "minimize_fleet_downtime": true,
    "consolidate_technician_visits": true,
    "respect_cleaning_schedules": true
  }
}

6.10 Security and Encryption

MQTT communication requires robust security to prevent unauthorized robot control, data interception, or denial-of-service attacks.

TLS Encryption

All MQTT connections use TLS 1.3 with mutual authentication. Robots present client certificates verifying their identity, while servers present certificates proving legitimacy.

MQTT Connection:
- Protocol: MQTT 5.0 over TLS 1.3
- Client Certificate: Robot-specific X.509 certificate
- Server Certificate: Broker CA-signed certificate
- Cipher Suite: TLS_AES_256_GCM_SHA384
- Perfect Forward Secrecy: Enabled

Topic-Level Access Control

MQTT brokers enforce fine-grained access control lists (ACLs):

Client Type	Publish Topics	Subscribe Topics
Robot	Own status/telemetry/alerts	Own commands, fleet messages
Fleet Controller	Fleet commands, task assignments	All robot status/telemetry
Edge Server	Model updates, path responses	Path requests, health data
Monitoring Dashboard	None	All status (read-only)

Review Questions

1. Why is publish-subscribe messaging better than request-response for multi-robot coordination?
2. Explain the difference between MQTT QoS levels 0, 1, and 2.
3. How does the topic hierarchy enable flexible subscription patterns?
4. Describe the collision avoidance protocol when robot trajectories intersect.
5. What triggers dynamic task reallocation in fleet operations?
6. How does edge computing enhance robot capabilities beyond on-device processing?
7. What are the three 5G network slice types and their robotics applications?
8. Explain how virtual pheromones optimize coverage in swarm cleaning.
9. Why is predictive maintenance more effective than reactive maintenance?
10. What security mechanisms protect MQTT communications from unauthorized access?