Micro, Lightweight, Intelligent Aerial Vehicles — a current scientific review (2025)

Authors: Anonymous review-style article · October 13, 2025

Abstract
Small, lightweight micro aerial vehicles (MAVs) have transitioned from experimental hobbyist platforms into a rapidly maturing class of intelligent robotic aircraft used in inspection, first response, ecology, precision agriculture, defense, and indoor logistics. This review summarizes the current (2024–2025) technological state of ultra-small MAVs (centimeter scale to small consumer drones), catalogs their features and capabilities, compares them to earlier generations, and highlights the principal advantages and remaining challenges.

Contents 1. Introduction — scope and definitions 2. Technological building blocks (current state) 3. Comprehensive list of features & capabilities (2024–2025 generation) 4. Comparison with legacy generations (pre-2015 / 2015–2020) 5. Advantages of the new generation 6. Outstanding challenges & research directions 7. Conclusion References

1. Introduction — scope and definitions

The term Micro Aerial Vehicle (MAV) typically denotes flying robots on the order of centimeters to tens of centimeters in characteristic size and from a few grams to a few kilograms in mass. This review focuses on the small and ultra-light end where weight, energy, and sensing/compute trade-offs are most acute. Included are: (a) centimeter-scale flapping/bio-inspired craft (gram-scale), (b) centimeter-to-20 cm multirotor platforms (tens to hundreds of grams), and (c) small lightweight quadcopters up to ~2–3 kg when constrained by mission.

2. Technological building blocks (current state)

2.1 Airframes & aerodynamics

Bio-inspired flapping wings, morphing surfaces, and highly optimized multirotor airframes reduce drag and increase lift-to-weight ratios for very small Reynolds numbers. Topology optimization and micro-composites are used to shave grams while preserving stiffness.

2.2 Propulsion & actuation

High-efficiency brushless coreless motors, specialized micro-propellers, and emerging electrohydrodynamic thrusters for very small craft permit higher propulsive efficiency at small scale. Design co-optimizes motor, propeller geometry, and control bandwidth.

2.3 Power & energy storage

Energy density, fast charge, and safety remain limiting factors. Advances in bespoke cells for drones (including pilot production of higher-density chemistries and early solid-state prototypes), plus fast-swap battery architectures and wireless/resonant charging, improve mission flexibility and endurance.

2.4 Sensing & perception

Lightweight LiDAR (MEMS), event cameras, efficient rolling-shutter sensors, ultrasonic/IR proximity arrays and integrated IMUs enable robust SLAM and obstacle avoidance even on gram-scale vehicles. Sensor fusion is commonly executed on-board.

2.5 On-board compute & machine learning

Edge AI accelerators (tiny NPUs), model pruning and quantization, and energy-aware learning enable on-device perception, policy inference and short-horizon planning. This reduces latency and ground-link dependence.

2.6 Communications & networking

Low-power mesh radios (sub-GHz, UWB, Wi-Fi variants) and ad-hoc networking protocols support local coordination. Federated learning and decentralized update strategies help conserve bandwidth and preserve fleet privacy.

3. Comprehensive list of features & capabilities (2024–2025 generation)

The list below captures capabilities available today or achievable with current commercial / near-commercial components (2024–2025):

• Autonomous navigation: vision + inertial SLAM, GNSS/RTK fallback, and indoor/outdoor transition handling.
• Obstacle avoidance: real-time reactive collision avoidance using vision, event cameras, LiDAR/ToF and ultrasonic sensors.
• Perception tasks: on-board object detection and semantic segmentation (e.g., crop/weed, person, structural defects) via quantized neural networks.
• Autonomous mission planning: short-horizon on-device planning with higher-level mission goals provided by operator or other agents.
• Agentic behaviors: goal-directed, memory-bearing behaviors enabling context-aware missions and opportunistic re-planning.
• Swarming & cooperation: distributed consensus, formation control, cooperative mapping and multi-agent sensing fusion.
• Edge learning & adaptation: local model updates, online domain adaptation, and federated learning for fleet improvements.
• Low SWaP sensors: MEMS LiDAR, event cameras, micro-IMUs, and thermal micro-cameras for tiny platforms.
• Energy management: intelligent power budgeting, battery swap capability, and wireless charging solutions.
• Payload versatility: micro-payload bays for environmental sensors, small manipulators, microphones, or tethered power/data links.
• Resilience features: fail-safe auto-land, lightweight parachutes/tethers in risky contexts, and encrypted communications.

4. Comparison with legacy generations (pre-2015 / 2015–2020)

Key differences

• Compute & autonomy: Legacy MAVs relied on ground-side processing and human teleoperation; modern MAVs run optimized models and planners on-board enabling near-full autonomy.
• Energy & endurance: Modern cells and propulsion improve flight time and payload capacity compared to older Li-ion-limited designs.
• Sensing sophistication: Multimodal sensing (MEMS LiDAR, event cameras, thermal) increases robustness in low-light and cluttered environments vs older monocular+IMU stacks.
• Cooperation: Swarm research moved from centralized/simulated proofs to decentralized real-world cooperative deployments.
• Manufacturing: Additive microfabrication and composite microstructures produce lighter, more integrated airframes than hobbyist-era designs.

Tabular comparison (high-level)

Aspect	Legacy MAVs (≤2015–2020)	Current MAVs (2024–2025)
Autonomy	Teleoperation or ground-side processing	On-board SLAM, agentic autonomy, on-device ML
Energy	Standard Li-ion, limited cycles	Higher-density cells, fast-swap, early solid-state pilots
Sensing	Monocular/stereo camera + IMU	MEMS LiDAR, event cameras, thermal, multimodal fusion
Networking	Point-to-point radio	Mesh, federated learning, secure comms
Swarm capability	Small-scale proofs, centralized concepts	Decentralized swarms, real-world cooperative strategies
Manufacturing	Commercial off-the-shelf, hobbyist	Microfabrication, composites, integrated sensors

5. Advantages of the new generation (concise)

1. Higher capability-to-weight ratio — more sensors, compute and payload per gram.
2. Reduced operator load — agentic autonomy lowers need for skilled pilots.
3. Improved mission endurance & flexibility — energy and propulsion advances yield longer flights.
4. Scalable cooperation — decentralized swarms enable complex collaborative tasks.
5. Privacy & resilience — edge processing reduces raw-data transmission and improves robustness.

6. Outstanding challenges & research directions

• Energy density vs. safety trade-offs: Advanced chemistries (solid-state) must mature for safe, mass-produced use.
• Perception robustness: Small sensors face motion blur, vibration, and low-light failure modes; fusion and event cameras help but are not panaceas.
• Certification & airspace integration: Regulation lags behind autonomous swarm capabilities and agentic behaviors.
• Secure low-latency networking: Mesh and federated systems need standardization and hardened implementations for large-scale deployments.

7. Conclusion

Between 2020 and 2025 the field of small, lightweight intelligent MAVs moved from constrained, human-in-the-loop platforms to fleet-capable, agentic systems with rich on-board perception and cooperative behaviors. Improvements in edge compute, sensing, networking and battery technology are the engine of this change. While materials, certification, and safety challenges remain, the new generation delivers markedly better autonomy, endurance, and multi-agent capability versus legacy MAVs — enabling applications ranging from autonomous indoor inspection to large-scale cooperative environmental sensing and resilient first-responder support.

Selected references (representative, non-exhaustive)

1. Recent state-of-the-art review articles on UAVs and AI (2025).
2. Survey on agentic UAVs integrating perception, planning and memory (arXiv / conference surveys, 2024–2025).
3. Frontiers review on MAV swarming and cooperative robotics (2020).
4. Industry reporting and announcements on advanced battery chemistries and drone-specific cells (2024–2025).

Note: This page is a generated review summarizing developments up to October 13, 2025. For exact citations, DOI links, or direct quotes, please provide a citation style or let me fetch and format primary sources.

If you want this exported as a PDF or a slide deck (PowerPoint), or if you want a formatted journal-style manuscript (with full bibliography and figure placeholders), tell me which format and I will generate it.