TAK Server MCP

import { TAKTool, ToolContext } from '../registry'; import * as turf from '@turf/turf'; export const analyzeMovementTool: TAKTool = { name: 'tak_analyze_movement', description: 'Track and analyze entity movements, patterns, and anomalies', category: 'geospatial', requiresAuth: true, requiresWrite: false, inputSchema: { type: 'object', properties: { entityId: { type: 'string', description: 'Entity ID to analyze movement for' }, timeRange: { type: 'object', properties: { startTime: { type: 'string', description: 'ISO 8601 date string - start of analysis period' }, endTime: { type: 'string', description: 'ISO 8601 date string - end of analysis period (defaults to now)' } }, description: 'Time range for movement analysis' }, analysisType: { type: 'array', items: { type: 'string', enum: ['speed', 'pattern', 'anomaly', 'prediction', 'stops'] }, default: ['speed', 'pattern'], description: 'Types of analysis to perform' }, anomalyThreshold: { type: 'number', minimum: 0, maximum: 1, default: 0.8, description: 'Threshold for anomaly detection (0-1)' } }, required: ['entityId'] }, handler: async (context: ToolContext) => { const { takClient, params, logger } = context; try { // Get CoT events for the entity within time range const endTime = params.timeRange?.endTime || new Date().toISOString(); const startTime = params.timeRange?.startTime || new Date(Date.now() - 24 * 60 * 60 * 1000).toISOString(); logger.debug(`Fetching movement data for entity ${params.entityId}`); const events = await takClient.getCotEvents({ uids: [params.entityId], start: new Date(startTime), end: new Date(endTime) }); if (events.length < 2) { return { success: true, data: { entityId: params.entityId, message: 'Insufficient movement data for analysis', dataPoints: events.length }, metadata: { timestamp: new Date().toISOString(), source: 'tak-server' } }; } // Sort events by time events.sort((a, b) => new Date(a.time).getTime() - new Date(b.time).getTime()); // Extract positions and timestamps const positions = events.map(event => ({ time: new Date(event.time), lat: event.point.lat, lon: event.point.lon, alt: event.point.hae || 0 })); // Initialize analysis results const analysis: any = { entityId: params.entityId, timeRange: { start: startTime, end: endTime, duration: (new Date(endTime).getTime() - new Date(startTime).getTime()) / 1000 }, dataPoints: positions.length }; const analysisTypes = params.analysisType || ['speed', 'pattern']; // Speed analysis if (analysisTypes.includes('speed')) { const speeds: number[] = []; let totalDistance = 0; for (let i = 1; i < positions.length; i++) { const from = turf.point([positions[i-1].lon, positions[i-1].lat]); const to = turf.point([positions[i].lon, positions[i].lat]); const distance = turf.distance(from, to, { units: 'meters' }); const timeDiff = (positions[i].time.getTime() - positions[i-1].time.getTime()) / 1000; if (timeDiff > 0) { const speed = (distance / timeDiff) * 3.6; // Convert to km/h speeds.push(speed); totalDistance += distance; } } analysis.speedAnalysis = { averageSpeed: speeds.length > 0 ? speeds.reduce((a, b) => a + b, 0) / speeds.length : 0, maxSpeed: speeds.length > 0 ? Math.max(...speeds) : 0, minSpeed: speeds.length > 0 ? Math.min(...speeds) : 0, totalDistance: totalDistance, units: { speed: 'km/h', distance: 'meters' } }; } // Pattern analysis if (analysisTypes.includes('pattern')) { // Calculate heading changes const headings: number[] = []; for (let i = 1; i < positions.length; i++) { const from = turf.point([positions[i-1].lon, positions[i-1].lat]); const to = turf.point([positions[i].lon, positions[i].lat]); const bearing = turf.bearing(from, to); headings.push(bearing); } // Detect circular patterns const headingChanges = []; for (let i = 1; i < headings.length; i++) { let change = headings[i] - headings[i-1]; if (change > 180) change -= 360; if (change < -180) change += 360; headingChanges.push(change); } const totalHeadingChange = headingChanges.reduce((a, b) => a + b, 0); const isCircular = Math.abs(totalHeadingChange) > 300; // Nearly full circle // Create movement path const coordinates = positions.map(p => [p.lon, p.lat]); const path = turf.lineString(coordinates); analysis.patternAnalysis = { movementType: isCircular ? 'circular' : 'linear', totalHeadingChange: totalHeadingChange, pathLength: turf.length(path, { units: 'meters' }) * 1000, boundingBox: turf.bbox(path) }; } // Stop detection if (analysisTypes.includes('stops')) { const stops = []; const stopThreshold = 50; // meters const stopDuration = 300; // 5 minutes let currentStop: any = null; for (let i = 1; i < positions.length; i++) { const from = turf.point([positions[i-1].lon, positions[i-1].lat]); const to = turf.point([positions[i].lon, positions[i].lat]); const distance = turf.distance(from, to, { units: 'meters' }); if (distance < stopThreshold) { if (!currentStop) { currentStop = { startTime: positions[i-1].time, location: positions[i-1], positions: [positions[i-1]] }; } currentStop.positions.push(positions[i]); currentStop.endTime = positions[i].time; } else if (currentStop) { const duration = (currentStop.endTime.getTime() - currentStop.startTime.getTime()) / 1000; if (duration >= stopDuration) { stops.push({ ...currentStop, duration: duration, center: turf.center(turf.points(currentStop.positions.map((p: any) => [p.lon, p.lat]))) }); } currentStop = null; } } analysis.stopAnalysis = { stops: stops, totalStops: stops.length, totalStopTime: stops.reduce((sum, stop) => sum + stop.duration, 0) }; } // Anomaly detection if (analysisTypes.includes('anomaly')) { const anomalies = []; // Detect speed anomalies if (analysis.speedAnalysis) { const avgSpeed = analysis.speedAnalysis.averageSpeed; const speedThreshold = avgSpeed * (1 + params.anomalyThreshold); for (let i = 1; i < positions.length; i++) { const from = turf.point([positions[i-1].lon, positions[i-1].lat]); const to = turf.point([positions[i].lon, positions[i].lat]); const distance = turf.distance(from, to, { units: 'meters' }); const timeDiff = (positions[i].time.getTime() - positions[i-1].time.getTime()) / 1000; if (timeDiff > 0) { const speed = (distance / timeDiff) * 3.6; if (speed > speedThreshold) { anomalies.push({ type: 'speed', time: positions[i].time, location: positions[i], value: speed, threshold: speedThreshold }); } } } } analysis.anomalyAnalysis = { anomalies: anomalies, totalAnomalies: anomalies.length }; } logger.info(`Movement analysis completed for entity ${params.entityId}`); return { success: true, data: analysis, metadata: { timestamp: new Date().toISOString(), source: 'tak-server' } }; } catch (error) { logger.error('Failed to analyze movement:', error); return { success: false, error: { code: 'TAK_ANALYSIS_ERROR', message: 'Failed to analyze entity movement', details: error instanceof Error ? error.message : 'Unknown error' } }; } } };