Skip to content

Future Enhancements & Roadmap

Table of Contents

  1. Overview
  2. Near-Term (6–12 Months)
  3. Mid-Term (12–24 Months)
  4. Long-Term (2+ Years)
  5. Roadmap Timeline

Overview

The Ramgarh Farm Sensor Node Rev A.1 is production-ready and field-proven. However, no design is perfect. Rev B and beyond will incorporate lessons learned from field deployment, improvements in component technology, and expanded capabilities. This roadmap outlines potential enhancements grouped by timeframe and effort level.

Enhancements are prioritized based on: 1. Impact on autonomy — Changes that extend battery life or reduce power consumption 2. Operational simplicity — Changes that reduce maintenance burden (fewer site visits, automatic configuration) 3. Data quality — Changes that improve measurement accuracy or temporal resolution 4. Cost — Changes that reduce BOM cost without sacrificing reliability 5. Risk reduction — Changes that address identified failure modes (e.g., MT3608 instability)

Near-Term (6–12 Months)

These improvements are low-hanging fruit, building on existing components and firmware.

1. Firmware Adaptive TX Interval

Current state: TX interval is fixed at 15 minutes (Profile B), regardless of conditions.

Enhancement: Firmware monitors WiFi signal quality, battery voltage, and solar harvest. It automatically adjusts TX interval: - Good WiFi (RSSI > -65 dBm) + strong solar → TX every 10 min (fresher data) - Marginal WiFi (-75 dBm) + weak solar → TX every 20 min (conserve energy, reduce retries) - Critical battery (<3.5V) → TX every 60 min (survival mode)

Benefit: - Maintains data freshness when conditions permit - Extends autonomy under degraded conditions (no manual profile switch needed) - Reduces operator burden (fully automatic)

Effort: Medium (firmware logic for signal detection and state machine) Impact: +10% autonomy on average (adaptive batching saves ~8 mAh/day)

2. On-Device Data Buffering to Flash

Current state: Data is buffered in RAM (3 readings, ~100 bytes). If WiFi fails, buffer overflows and old readings are lost.

Enhancement: Use ESP32 onboard SPI flash (typically 4 MB) for persistent data buffer. If WiFi is down for hours or days, keep 100+ readings in flash. Once WiFi recovers, transmit buffered data to server in batches.

Benefit: - No data loss during WiFi outages - Better trend analysis (filled-in gaps reduce interpolation artifacts) - Easier root-cause analysis (can see exactly when WiFi failed)

Effort: Medium (flash wear management, ring buffer implementation) Impact: Data integrity (not autonomy)

3. Battery Health Prediction

Current state: Operator must manually estimate battery age and schedule replacement at 18 months.

Enhancement: Firmware monitors battery voltage trend. Once per week, it performs an open-circuit voltage (OCV) measurement (power off for 1 minute, measure voltage). OCV is a proxy for state-of-charge. Plot OCV over months to detect capacity degradation trend.

Benefit: - Automatic prediction of battery end-of-life - Alert sent to operator 4 weeks before expected failure - Data-driven replacement schedule (replace when needed, not on fixed schedule)

Effort: Medium (requires OCV calibration curve for NCR18650B, trend analysis firmware) Impact: Reduced maintenance (fewer surprised failures)

4. Enclosure Temperature Monitoring

Current state: No internal temperature measurement. Thermal stress is invisible.

Enhancement: Add DHT22 on GPIO16 (optional, populated by default in Rev B). Log enclosure temperature daily. If >55°C, firmware alerts and automatically switches to Profile C (lower consumption = less heat).

Benefit: - Detects overheating problems early - Automatic power reduction prevents thermal runaway - Data helps validate thermal model (compare actual to IPC-2221 prediction)

Effort: Low (DHT22 is optional, already in current design) Impact: Thermal stability (prevents failures in extreme heat)

5. Remote Configuration Portal

Current state: Firmware configuration requires USB connection and re-flash.

Enhancement: Build a simple web interface accessible via farm server. Operator can: - Change WiFi SSID and password - Switch profiles (A, B, C) - Set TX interval - View live battery voltage and WiFi signal - Trigger manual data transmission

Benefit: - No USB cable or laptop needed in the field - Configuration changes take effect immediately (no reboot needed) - Remote diagnostics possible (check WiFi strength from office)

Effort: Medium (requires HTTP/JSON server on node, or cloud API integration) Impact: Operational simplicity (fewer field visits)

Mid-Term (12–24 Months)

These enhancements require component changes or significant firmware redesign.

1. Improved Boost Converter (Lower Quiescent Current)

Current state: MT3608 draws 0.6 mA quiescent, dominates sleep current (79% of total).

Enhancement: Replace MT3608 with TPS63001 or similar ultra-low-quiescent buck-boost IC (~0.1 mA quiescent). Smaller SMD package, may require PCB redesign.

Benefit: - Sleep current reduced from 0.76 mA to ~0.25 mA - Daily consumption drops from 77.6 mAh to ~50 mAh - Autonomy increases from 31 days to 54 days (no solar) - 4.5× solar margin becomes 7.0× (much more resilient)

Effort: High (component change, PCB redesign, breadboard validation) Impact: Major (14% reduction in daily consumption, 23-day increase in autonomy)

Risks: New IC may have different stability characteristics; must validate on breadboard.

2. MPPT Solar Charging

Current state: MCP73871 is a linear charger (constant-current, no maximum power point tracking).

Enhancement: Add a dedicated MPPT (Maximum Power Point Tracking) IC that monitors panel voltage/current and dynamically adjusts load to extract maximum power. Increases effective harvest by 20–30% on non-ideal conditions (haze, dust, temperature).

Benefit: - 20–30% increase in harvested energy on cloudy days - System more resilient to dust and degradation - Margin increases from 4.5× to 6.0× or better

Effort: High (MPPT IC selection, firmware integration, efficiency validation) Impact: Significant (extends autonomy by ~5 days on cloudy days)

Risks: Cost increases (~$20 per node); complexity increases.

3. Dual-Supply Architecture (3.3V Sensors + WiFi on 5V)

Current state: All sensors are powered from 3.3V LDO (which draws 1 mA quiescent).

Enhancement: Split supplies: sensors on 3.3V (powered by battery directly, eliminating LDO), WiFi on 5V (boosted from battery). Requires buck converter for 3.3V, saving LDO power.

Benefit: - Eliminate quiescent draw from AMS1117 LDO (~1 mA) - Sleep current reduced by ~15% - More flexible power management (can power-gate sensors independently)

Effort: High (new buck IC, PCB redesign, firmware changes) Impact: Moderate (~5% autonomy improvement)

Risks: Added component count; more complex power sequencing.

4. 8×8 Grid Support (Expanded Sensor Density)

Current state: Rev A supports single node per location; digital twin has 4×4 grid of zones.

Enhancement: Firmware supports daisy-chaining multiple nodes via shared WiFi or mesh network. Backend server aggregates data from up to 64 nodes (8×8 grid). PCB form factor remains unchanged; multiple nodes are deployed across farm.

Benefit: - Higher spatial resolution (crop moisture at finer granularity) - Better trend detection (interpolation between nodes more accurate) - Scalability to larger farms

Effort: High (server architecture changes, mesh networking firmware) Impact: Major (unlocks 8×8 grid capability)

Risks: Server and networking infrastructure must support increased data rate.

5. OTA Firmware Update Infrastructure

Current state: Firmware updates require USB cable and re-flash.

Enhancement: Implement firmware update-over-the-air. Node fetches new firmware binary from server via WiFi, verifies signature, writes to flash, and reboots.

Benefit: - Deploy bug fixes and feature updates without field visits - Rapid response to discovered issues - Automatic rollout to fleet

Effort: High (secure bootloader, signature verification, rollback handling) Impact: Operational simplicity (no more field visits for firmware)

Risks: Security (firmware tampering, improper update could brick node); requires careful testing.

Long-Term (2+ Years)

These are aspirational improvements, dependent on technology roadmap and farm requirements.

1. Multi-Sensor Pod Clustering

Current state: Each node has fixed sensor suite (soil, rain, temperature, light).

Enhancement: Design modular sensor pods that can be swapped: - Crop pod: Crop reflectance (NDVI), stem water potential, leaf moisture - Soil pod: Extended soil profile (multiple depths), soil matric potential - Microclimate pod: Wind speed, detailed radiation, CO₂

Each pod has 4-pin connector to central node. Operator can swap pods seasonally.

Benefit: - Single node hardware supports diverse monitoring strategies - Modularity reduces BOM complexity - Future-proofs the system for new sensors

Effort: Very high (mechanical design, sensor calibration, firmware abstraction) Impact: Major (enables precision agriculture use cases)

2. Camera Integration

Current state: No visual monitoring.

Enhancement: Add low-power 2MP camera module (OV2640) on optional connector. Firmware captures a daily crop photograph and transmits thumbnail to server (compression reduces from ~250 KB to ~10 KB).

Benefit: - Visual crop monitoring (canopy color, weed pressure, pest damage) - Temporal image mosaic (time-lapse animation of growth) - Easier diagnosis of field problems

Effort: Very high (low-power camera control, JPEG compression, server storage) Impact: Major (adds visual layer to crop monitoring)

Risks: Data rate and storage increase 100×; power consumption may exceed budget (requires camera gating).

3. Distributed Mesh Network

Current state: WiFi assumes centralized access point on farm.

Enhancement: Nodes form a LoRaWAN or Bluetooth Mesh network. Data relayed hop-by-hop from nodes to a central gateway. Increases range (can reach distant fields), fault tolerance (nodes help each other, network healing).

Benefit: - Range limited only by number of hops, not WiFi radius - Self-healing (mesh routing around failed nodes) - Supports 100+ nodes efficiently

Effort: Very high (mesh protocol, routing, interference management) Impact: Major (unlocks large-scale farm monitoring, 8×8 grid becomes feasible)

Risks: Complexity; LoRa transceivers add cost (~$15 per node); requires careful frequency planning (ISM band licensing).

4. AI-Powered Anomaly Detection

Current state: Farm server stores raw data; operator reviews dashboards.

Enhancement: Deploy a machine learning model on the farm server (or edge device) that learns normal patterns in soil moisture, temperature, and humidity. Automatically flags anomalies: - Sudden moisture drop (irrigation valve failure?) - Temperature spike (fire? equipment malfunction?) - Unexplained humidity surge (water leak? pest migration?)

Benefit: - Early warning system for problems - Reduces manual monitoring effort - Enables predictive maintenance

Effort: High (model training, labeled data collection, inference pipeline) Impact: Major (shifts from reactive to proactive monitoring)

Risks: Requires 6–12 months of labeled training data; false positives can desensitize operator.

Current state: Single WiFi link; no fallback.

Enhancement: Add sub-GHz radio (433 MHz or 915 MHz, LoRa or LoRaWAN) as fallback. If WiFi is unavailable for >1 hour, switch to LoRa to transmit critical data (battery status, alarm).

Benefit: - Improved coverage in dense vegetation or valleys - Graceful degradation (reduced data rate, not total loss) - Enables remote monitoring without infrastructure

Effort: High (second radio, antenna switching, firmware state machine) Impact: Moderate (improves reliability, not autonomy)

Roadmap Timeline

timeline
    title Sensor Node Enhancement Roadmap

    section Rev A.1 (Production)
        2026-Q1 : Breadboard validation complete
        2026-Q1 : PCB fabrication & assembly
        2026-Q2 : Field trial (4 nodes, 3 zones)
        2026-Q3 : Production deployment (16 nodes, full 4x4 grid)

    section Rev B (12 Months)
        2026-Q4 : Analyze field data, identify issues
        2027-Q1 : Design improvements (MPPT, lower quiescent boost)
        2027-Q2 : Firmware enhancements (adaptive TX, flash buffering)
        2027-Q3 : Commissioning & validation
        2027-Q4 : Begin Rev B field deployment (mixed fleet A+B)

    section Rev C (24 Months)
        2027-Q4 : Plan 8x8 grid scaling
        2028-Q1 : Design mesh networking stack
        2028-Q2 : OTA firmware infrastructure
        2028-Q3 : Sensor pod modularity
        2028-Q4 : Begin Rev C pilot (small mesh network)

    section Advanced (3+ Years)
        2029-Q1 : Camera integration prototype
        2029-Q2 : AI anomaly detection research
        2029-Q3 : Multi-spectral sensor evaluation
        2029-Q4 : Vision for next-generation digital twin

Decision Gate: When to Revisit

Each enhancement should be reconsidered after field data is available. Decision criteria:

Proceed with enhancement if: - ✓ Field data validates model assumptions (measured autonomy matches prediction within ±15%) - ✓ No critical failures in first 6 months - ✓ Operator feedback prioritizes this enhancement - ✓ Cost-benefit analysis is positive (cost < value to farm)

Defer enhancement if: - ✗ Field results show different priorities (e.g., WiFi reliability matters more than autonomy) - ✗ Simpler alternatives emerge (e.g., better solar panel, not MPPT) - ✗ Component technology changes make current design obsolete (e.g., new micro-controller with better WiFi stack)

Next Steps

← Deployment Guide | Back to Overview