Beyond Obstacle Avoidance: What First Responders Must Know About Drone Operations Over People

Public safety agencies nationwide are rapidly adopting drones for crowd monitoring, emergency response, and situational awareness. These platforms provide valuable real-time intelligence that can improve officer safety and operational effectiveness. However, as drone operations over people become routine, agencies must carefully balance operational capabilities against physical risks, including impact injuries and lacerations. This article examines the physics of drone safety, analyzes lessons from field deployments including a recent near-miss incident, and offers evidence-based recommendations to help first responders optimize both effectiveness and public safety.

Understanding Impact Physics: Beyond Simple Calculations

When assessing drone safety over crowds, the fundamental consideration is potential impact energy and injury risk in the event of an unexpected fall. However, simple gravitational potential energy calculations (E = mgh) can significantly overestimate real-world risks by assuming vacuum conditions and rigid body impacts.

Research from institutions including Virginia Tech, the FAA’s ASSURE program, and studies published in the AIAA Journal and International Journal of Crashworthiness reveals that actual collision dynamics are far more complex. Key factors include:

Terminal Velocity Effects: Drones don’t free-fall—air resistance limits descent velocity to approximately 18-30 m/s for typical quadcopters depending on size, orientation, aerodynamic profile, and whether propellers are rotating. Free-fall tests have measured terminal velocities around 18.5 m/s, while aerodynamic modeling suggests values approaching 29-30 m/s for certain configurations. This substantially reduces kinetic energy compared to potential energy calculations, particularly for drops exceeding ~100 feet where terminal velocity is reached. Variability is significant: Propeller rotation or tumbling can further reduce effective velocity. Terminal velocity estimates assume stable descent orientation; actual values in tumbling or flat-spin scenarios may vary by ±30%.

Energy Dissipation and Frangibility: Drones typically deform, fragment, or tumble during falls, dispersing energy across multiple impact points rather than delivering concentrated force. Frangible designs (e.g., breakable components) can absorb 50-80% of kinetic energy, as noted in FAA ASSURE reports. The FAA recognizes this in Part 107 regulations, where frangible multirotors can qualify for operations over people despite higher pre-impact energy due to favorable energy transfer characteristics—for example, a DJI Phantom 3 can meet Category 2 requirements with up to ~130 ft-lb (~176 J) kinetic energy.

Laceration Risks: Beyond blunt trauma, exposed rotating parts (e.g., propellers) pose significant laceration hazards. FAA regulations emphasize designs that prevent skin laceration, with testing using anthropomorphic models to assess both kinetic energy transfer and cutting risks.

FAA Injury Thresholds and Categories: In Part 107 Subpart D, the FAA created different safety categories for flying drones over people. To figure out which drones belong in which category, they crash-tested drones into dummy heads and fake skin that’s designed to react like real human heads and skin. Based on how much damage the drones caused in these tests, they sorted them into groups—from ‘safest’ to ‘needs the most restrictions.:

• Category 1: Eligible for small unmanned aircraft weighing 0.55 pounds (250 grams) or less on takeoff (including all attachments). Must have no exposed rotating parts capable of lacerating human skin. No specific kinetic energy threshold, as safety is inherent to low weight and design.
• Category 2: Must be shown not to transfer more than 11 ft-lb (~15 J) of kinetic energy to a human upon impact when tested (equivalent to a rigid object). No exposed rotating parts capable of lacerating skin; no safety defects. Propellers must be enclosed or guarded. Sustained flight over open-air assemblies is restricted unless the drone complies with remote ID requirements.
• Category 3: Must be shown not to transfer more than 25 ft-lb (~34 J) of kinetic energy when tested. Similar laceration and defect prohibitions as Category 2. No operations over open-air assemblies of people. Sustained flight over people allowed only in closed or restricted-access areas, with notifications to participants or operations under protective structures/vehicles.
• Category 4: Requires an airworthiness certificate under Part 21. Allows operations over people if compliant with the certificate’s limitations, including maintenance logs retained for at least one year.

These thresholds focus on transferred kinetic energy (post-dissipation) rather than pre-impact values. Compliance often requires a Declaration of Compliance (DOC) or Means of Compliance (MOC) submitted to the FAA, demonstrating testing for energy transfer, laceration resistance, and absence of defects. In practice, factors like glancing blows, protective clothing, impact location, and individual physiology affect outcomes. FAA ASSURE research indicates injury probabilities vary: e.g., 11-13% chance of neck injury at 114-121 ft-lb transferred energy. General correlations suggest energies below 100 J often result in minor injuries (bruising, minor lacerations), while above 200-300 J increase risks of serious trauma (fractures, internal injuries), though lacerations can occur at lower energies if propellers are exposed.

Failure Rate Context: Reliability data varies by platform class, operating environment, and maintenance. A comprehensive review of small UAS operations suggests failure rates around 10⁻³ per flight hour (0.1%) across diverse platforms, though enterprise-grade systems with redundant components may achieve rates below 0.01%. Common failure modes include battery issues (~40% of incidents), software glitches, and environmental factors. Regular maintenance, redundant systems, and professional operation significantly improve reliability.

Additional Hazards: Beyond ground impacts, consider airborne risks such as mid-air collisions with birds, other aircraft, or objects like balloons, which can cause fragmentation and debris. Post-crash battery fires pose secondary hazards, particularly in crowded areas. Electromagnetic interference in urban environments can also disrupt operations. Psychological impacts on crowd behavior (panic response to perceived threats or falling drones) should also be considered in risk assessments and operational planning.

Comparative Analysis of Common Public Safety Drones

The following table compares various drone models, some of which are frequently used by first responders, showing both theoretical potential energy from a 200-foot altitude (typical for crowd monitoring) and more realistic terminal velocity-based kinetic energy estimates at 18-30 m/s. We have included flight-time and camera capabilities to show that smaller, lighter drones can provide similar crowd monitoring capabilities:

Drone	Weight	Potential Energy (J)	Terminal Velocity KE (J)*	Flight Time	Camera Capabilities
DJI Mini 5 Pro	0.253 kg / 0.55 lbs	150	40-110	36 min	1″ CMOS, 50MP, 4K/60fps, 2× digital
DJI Air 3S	0.724 kg / 1.60 lbs	433	120-330	45 min	Dual 1″+1/1.3″ CMOS, 48MP, 4K/120fps, 3× optical
DJI Matrice 4	1.219 kg / 2.69 lbs	728	200-550	49 min	Tri-camera: 48MP wide/70mm/168mm + 640×512 thermal, laser rangefinder
Skydio X10	2.11 kg / 4.65 lbs	1,263	340-950	40 min	Dual 1″+1/1.7″ CMOS, 50.3+64MP, 4K/60fps, ~4× optical
DJI Matrice 30	4.069 kg / 8.97 lbs	2,430	660-1,830	~41 min**	Multiple payload options, 1″ CMOS visible light, 10× optical
DJI Matrice 400	15.8 kg / 34.8 lbs	9,441	2,560-7,110	~35 min**	High-end multisensor payloads, 4/3″+additional sensors, 30× optical

*Terminal velocity KE calculated using v = 18-30 m/s based on aerodynamic studies. Actual impact energy is orientation-dependent; tumbling or horizontal descent reduces effective velocity. Frangible construction and propeller/structure deformation absorb significant energy, which the FAA considers in Operations Over People equivalence determinations. Note that sub-250g drones like the Mini 5 Pro may avoid some registration requirements but can still exceed limits due to manufacturing variances. These platforms may not be certified for all Part 107 categories; operators should verify specific model certifications and compliance documentation with manufacturers.

**Flight times listed by manufacturers. These times can vary in different weather conditions and can differ significantly with payload configuration for enterprise models.

*** The DJI Mini 5 Pro, and the DJI Air 3S are very capable consumer drones and could be used in certain first responder applications, but they lack any kind of thermal camera.

*** This table does not include all currently available drones from all drone manufacturers in the United States. We have included a number of widely used drones to illustrate the safety considerations and aircraft capabilities.

This data demonstrates clear mass-risk correlation while acknowledging that modern safety systems—redundant batteries, AI obstacle avoidance, automatic return-to-home, geofencing—substantially reduce failure probability compared to earlier drone generations. Agencies should also consider propeller cages, parachutes, integration costs, insurance requirements for over-people operations, and platform-specific advantages like wind resistance for heavier models.

Case Study: Learning from Real-World Deployments

NYPD’s Drone as First Responder Program

The New York Police Department has been at the forefront of law enforcement drone adoption. On September 23, 2024, the NYPD received FAA approval for Drone as First Responder (DFR) operations without visual observers, enabling autonomous deployment across New York City’s complex urban environment. The department has deployed Skydio X10 platforms for these operations, leveraging the platform’s advanced capabilities.

Skydio claims that the X10 platform offers several operational advantages for first responders:

• Modular multi-sensor payloads including high-resolution RGB and Boson+ thermal (or a dedicated low-light module).
• ~40-minute endurance; published LOS ranges up to 12 km in rural/clean-RF conditions, with 5G backhaul where available.
• Designed and assembled in the U.S.; X10D variant is on DoD’s Blue UAS Cleared List.
• IP55-rated airframe for light–moderate precipitation and dust; tested to ~28 mph gusts (avoid icing/lightning).

Recent deployments have included crowd monitoring at public events and demonstrations, suspect tracking operations, and emergency response coordination. The department reports no injuries from drone operations to date, demonstrating the effectiveness of professional piloting and safety protocols.

Operational Considerations: Autonomy vs. Manual Control

While enterprise drones tout advanced autonomous features and AI-powered obstacle avoidance as operational advantages, field experience reveals important nuances that agencies should consider when selecting platforms.

The Manual Control Preference:

Many experienced pilots, particularly those with traditional aviation backgrounds, prefer aircraft that respond predictably to direct control inputs. During time-critical operations—suspect tracking, search and rescue, or tactical support—pilots want immediate, precise aircraft response without AI intervention. Some autonomous systems, while technologically sophisticated, can introduce latency or unexpected behavior that complicates precision maneuvering.

Autonomy Trade-offs in DFR Operations:

During Drone as First Responder deployments, several operational realities emerge:

• Geofencing vs. Obstacle Avoidance: Agencies typically establish geofenced authorization zones that define where aircraft can operate. In these pre-cleared operational areas, active obstacle avoidance may provide limited additional value while potentially interfering with pilot control authority.
• AI Control Authority: Some platforms with aggressive AI-assisted flight can override pilot inputs to avoid perceived obstacles. While this may prevent collisions in some scenarios, it can also frustrate pilots attempting precision maneuvers, particularly in confined spaces or time-sensitive situations.
• Performance in Standardized Testing: Operational testing has revealed performance variations across platforms. Some AI-assisted drones receive time forgiveness penalties in standardized precision courses (such as NIST evaluation protocols) due to less predictable control response, reflecting the inherent tension between autonomous assistance and direct manual control.

Camera System Performance:

Beyond flight characteristics, camera quality remains critical for first responder operations. While manufacturers publish impressive sensor specifications, real-world performance—particularly in challenging lighting, at maximum zoom, or during dynamic tracking—varies significantly between platforms. Agencies should conduct extensive field evaluations rather than relying solely on technical specifications, with particular attention to:

• Image quality at operational zoom levels (not just wide-angle performance)
• Low-light and high-contrast performance
• Stabilization effectiveness during movement
• Thermal/visible sensor integration (where applicable)

Recommendations:

1. Conduct Hands-On Evaluations: Before committing to a platform, ensure pilots who will operate the aircraft daily have extensive hands-on time with candidate systems in realistic operational scenarios.
2. Prioritize Pilot Preference: Technology sophistication matters less than whether your pilots can effectively operate the platform under pressure. A less autonomous drone that responds predictably may outperform a more advanced system that fights pilot inputs.
3. Test in Operational Contexts: Evaluate platforms during scenarios matching your actual use cases—not just manufacturer demonstrations. Test precision maneuvering, time-to-target, and camera performance in your specific environment.
4. Consider Control Philosophy: Determine whether your agency values autonomous assistance or direct manual control, and select platforms aligned with that operational philosophy. Both approaches have merit; the key is matching technology to pilot preference and mission requirements.

Different manufacturers emphasize different control philosophies—some prioritizing autonomous assistance, others emphasizing responsive manual control. Neither approach is inherently superior; effectiveness depends on operational context, pilot experience, and mission requirements. Agencies should prioritize platforms their pilots can operate confidently and effectively, regardless of marketing claims about technological sophistication.

The Balloon Near-Miss Incident: Understanding Obstacle Avoidance Limitations

During an October 2025 public event, an NYPD Skydio X10 operating at approximately 200 feet altitude over a crowd came into close proximity with a helium balloon. While the balloon and drone did not collide, it came close enough to the drone to highlight a critical operational consideration: the inherent limitations of vision-based obstacle avoidance systems in real-world urban environments.

Understanding AI Obstacle Avoidance Capabilities and Constraints:

Modern enterprise drones like the Skydio X10 employ sophisticated AI-powered obstacle avoidance using vision sensors and machine learning algorithms. These systems excel at detecting and avoiding stationary objects with clear visual contrast—buildings, trees, vehicles, and other fixed structures. However, all vision-based obstacle avoidance systems, regardless of manufacturer, face significant technical challenges with:

Moving Objects: AI obstacle avoidance is optimized for static environments. Moving objects—including balloons drifting in wind, birds in flight, or other aircraft—present substantially greater detection and avoidance challenges. The system must not only detect the object but predict its trajectory and calculate an avoidance path in real-time, which becomes exponentially more complex with unpredictable movement patterns.

Thin Obstacles: Balloon strings, power lines, thin branches, and wires present particular difficulty for vision-based detection systems. These objects may occupy only a few pixels in the camera’s field of view and can be easily lost against complex backgrounds or in certain lighting conditions. Even advanced machine learning algorithms struggle to reliably identify and track such minimal visual signatures.

Environmental Conditions: Vision-based obstacle avoidance performance degrades significantly under challenging conditions including:

• Direct sunlight or backlighting: Can cause lens flare, sensor saturation, or silhouetting that obscures obstacles
• High contrast scenarios: Bright sky backgrounds make thin, translucent, or reflective objects (like balloon strings) nearly invisible
• Weather conditions: Rain, fog, snow, or dust reduce sensor range and image clarity
• Low light: Diminishes the visual information available for object recognition, even with infrared sensors

What this incident demonstrates:

• Urban environments contain dynamic, unpredictable airborne hazards that challenge even the most advanced obstacle avoidance technology
• Vision-based systems—while highly capable—have inherent physical and computational limitations that operators must understand
• Environmental factors like sunlight, weather, and background contrast significantly affect detection reliability
• Thin, moving objects like balloon strings represent a worst-case detection scenario for current technology
• The importance of comprehensive environmental hazard assessment in pre-flight planning
• Why human pilot judgment and situational awareness remain critical safety layers

What this incident does NOT necessarily indicate:

• A failure of Skydio technology or its obstacle avoidance algorithms
• Inadequate pilot training or decision-making by the NYPD operator
• Unacceptable risk in properly conducted operations
• A deficiency unique to any particular manufacturer

Rather, this near-miss should be understood as a valuable learning opportunity that underscores the fundamental limitations that affect all vision-based obstacle avoidance systems across the industry. Similar scenarios would challenge any drone manufacturer’s platform operating under the same environmental conditions. This is about the current state of sensor technology and physics, not about specific equipment design or operator performance.

The incident reinforces why drone operations over people require multiple overlapping safety measures: robust technology, skilled pilots, visual observers, environmental assessment, appropriate altitude selection, and—when feasible—lateral offset from crowds. No single layer of protection is sufficient in complex urban environments.

Evidence-Based Recommendations for Safer Operations

1. Mission-Appropriate Platform Selection

Principle: Use the lightest drone that meets genuine operational requirements.

Implementation:

• Routine crowd monitoring: Consider lighter platforms (DJI Mini 5 Pro, Air 3S) when standard 4K video quality and basic zoom capabilities suffice
• Enhanced intelligence gathering: Mid-weight platforms (DJI Matrice 4, Skydio X10) when thermal imaging, AI tracking, or adverse weather capability is operationally justified
• Specialized operations only: Heavy enterprise platforms (DJI Matrice 30/400) reserved for missions requiring specialized payloads or extreme weather tolerance

Risk-benefit analysis: A DJI Air 3S delivers 48MP imaging with 3× optical zoom at 724g—potentially adequate for many crowd monitoring scenarios while carrying 83% less mass than a 2.11kg Skydio X10. However, the X10’s autonomous navigation, superior obstacle avoidance, and dual thermal/visible sensors may prove essential for complex urban operations or low-light scenarios, justifying the additional mass.

2. Flight Parameter Optimization

Altitude management: Impact energy scales with drop height, though the relationship is non-linear once terminal velocity is approached. For drones that reach terminal velocity quickly (lighter models), reducing altitude from 200 feet to 100 feet may reduce impact energy by less than 50% since velocity is already capped. However, minimizing altitude when mission allows remains a best practice for all platforms.

Lateral offset strategy: Position drones adjacent to rather than directly above crowds whenever operationally feasible. A 30-50 foot lateral offset maintains visual coverage while dramatically reducing direct strike probability.

Temporal minimization: Reduce overhead flight time to mission-essential duration. Brief surveillance passes may accomplish objectives while limiting risk exposure.

2.5. Critical: Minimize Overhead Flight Duration

One of the most significant risk-multiplying factors is sustained hovering directly over crowds. During the October 2025 observation, NYPD Skydio X10 drones maintained a fixed position approximately 200 feet over protesters for extended periods—hovering until battery depletion required landing, then immediately relaunching to resume the same overhead position. This approach, while maximizing continuous surveillance coverage, substantially increases cumulative risk exposure.

Why sustained overhead hovering multiplies risk:

• Extended exposure window: A drone hovering for 30-40 minutes creates hundreds of times more exposure than a 30-second pass
• Fatigue and complacency: Prolonged operations increase pilot fatigue and can reduce vigilance for emerging hazards
• Static positioning: Fixed-position hovering offers no margin for error—any failure results in direct vertical descent into the crowd below
• Battery depletion risk: Operating to near-battery-depletion before landing creates additional failure modes and reduces safety margins for emergency maneuvering

Risk calculation perspective: If a drone has a 0.1% per flight-hour failure probability, a 5-minute surveillance pass carries roughly 0.0083% risk exposure, while a 40-minute hover increases exposure to 0.067%—an 8× increase for the same intelligence gathering mission.

Best practices for temporal risk management:

1. Dynamic patrol patterns: Instead of fixed hovering, establish orbital or figure-eight patterns that keep the drone adjacent to rather than directly above crowds while maintaining observation angles
2. Segmented coverage: Divide monitoring area into zones; rotate between zones rather than maintaining continuous overhead presence in any single location
3. Time-limited overhead passes: When direct overhead positioning is operationally necessary, limit duration to mission-essential minimums (e.g., 2-5 minute passes rather than continuous hovering)
4. Battery management protocols: Establish return-to-home triggers at 30-35% remaining battery rather than flying to near-depletion, preserving energy reserves for emergency responses or unexpected wind conditions
5. Mission planning: Pre-flight analysis should identify whether continuous overhead surveillance is genuinely necessary or whether intermittent passes (every 5-10 minutes) would achieve operational objectives with 80-90% reduced exposure time
6. Alternative coverage methods: Consider whether elevated camera positions, multiple drones rotating coverage, or ground-based observation can reduce individual aircraft overhead duration

The fundamental principle: Overhead flight time is a direct multiplier of risk exposure. Every additional minute directly above people compounds the consequences of any potential failure. Agencies should actively minimize this exposure rather than treating extended overhead hovering as default operating procedure.

3. Enhanced Safety Systems

Hardware mitigations:

• Parachute systems that can serve as part of an FAA-accepted Means of Compliance (MOC) for Category 2/3 operations under Part 107 Subpart D
• Propeller guards where compatible with flight performance (required for certain categories)
• Redundant battery systems and failsafe protocols
• Enhanced GPS/IMU redundancy

Software protections:

• Geofencing to prevent unauthorized incursions over crowds
• Automatic altitude limitations
• Low-battery return-to-home with adequate margin
• Real-time telemetry monitoring

Agencies should consult Advisory Circular 107-2A for detailed guidance on acceptable means of compliance for operations over people.

4. Environmental Hazard Assessment

Given the inherent limitations of vision-based obstacle avoidance systems, comprehensive environmental assessment becomes critical:

Pre-flight evaluation:

• Airborne hazard survey: Identify balloon vendors, kite activity, decorative banners, and other floating objects. Note that even stationary balloons become moving hazards in wind.
• Thin obstacle mapping: Document power lines, cables, guy wires, and tree canopy areas with thin branches that may challenge visual detection.
• Lighting conditions: Assess sun angle and predict times when backlighting or direct sunlight may compromise obstacle detection. Consider rescheduling operations during suboptimal lighting windows.
• Weather assessment: Evaluate wind (affects balloon movement and drone stability), precipitation, fog, or atmospheric conditions that degrade sensor performance.
• Bird activity patterns: Research local bird migration seasons, nesting areas, and typical flight altitudes.
• Backup landing zones: Identify clear areas for emergency landing away from crowds.

Dynamic monitoring:

• Maintain visual observers during over-people operations where required, with specific instructions to monitor for airborne hazards
• Establish communication protocols for emerging hazards (e.g., balloon release, bird swarm)
• Define clear trigger criteria for operation suspension based on environmental changes
• Monitor real-time weather conditions and sensor performance indicators

Operational adjustments for high-risk conditions:

• Increase lateral offset from crowds when thin obstacles (strings, wires) are present
• Reduce flight speed in areas with unpredictable moving hazards
• Consider alternative monitoring methods (elevated camera positions, lighter tethered systems) when environmental conditions significantly degrade obstacle avoidance reliability

5. Operational Protocols and Training

Standard Operating Procedures (SOPs):

• Document risk assessment methodology aligned with Part 107 Subpart D categories
• Establish authorization levels for different drone weights over people
• Include specific procedures for obstacle avoidance system limitations and degraded sensor conditions
• Define emergency response procedures
• Create clear criteria for operation suspension

Pilot qualification and training:

• FAA Part 107 certification (minimum requirement)
• Advanced training specific to over-people operations
• Technical education on obstacle avoidance limitations: Ensure pilots understand the difference between stationary and moving object detection, recognize conditions that degrade sensor performance, and know how to compensate with manual inputs
• Regular scenario-based exercises including emergency procedures and sensor-degraded operations
• Ongoing education on new safety technologies and incident case studies

Documentation and review:

• Log all over-people operations with flight parameters and environmental conditions
• Conduct after-action reviews including near-miss analysis with focus on obstacle avoidance system performance
• Share lessons learned across agencies and with equipment manufacturers
• Update protocols based on incident data and evolving technology capabilities

6. Regulatory Compliance and Community Engagement

Regulatory Pathways for First Responders:

Public safety agencies have two primary pathways for drone operations:

• Part 107 Operations: Many routine flights can be conducted under Part 107 with appropriate declarations for over-people operations
• COA (Certificate of Waiver or Authorization): Government entities can apply for COAs that waive specific Part 107 restrictions for public safety missions. COAs commonly enable:
  • BVLOS operations for search and rescue or suspect tracking
  • Night operations without special lighting
  • Operations over people during emergency response
  • Higher altitude operations when needed

Important: Regardless of regulatory authorization (Part 107 or COA), the physics-based safety considerations, environmental limitations, and operational best practices discussed in this article apply equally to all operations.

FAA alignment:

• Operate under appropriate Part 107 categories for over-people operations (Categories 1-4)
• Maintain required declarations, declarations of compliance (DOC), or means of compliance (MOC) documentation
• Pursue beyond visual line of sight (BVLOS) waivers only with commensurate safety justifications
• Document compliance with operational limitations

Public transparency:

• Clearly communicate drone program objectives and benefits
• Explain safety measures including both technological capabilities and their limitations
• Address community concerns constructively, including privacy protections
• Publish aggregate operational statistics (flights, incidents, outcomes)

Manufacturer Considerations: What Industry Can Do

While this article focuses on operational practices, manufacturers play a crucial role in advancing safety:

Design innovations:

• Continue developing lighter materials, more efficient motors, and improved aerodynamics that reduce weight without sacrificing capability
• Advance sensor fusion techniques combining multiple detection methods (vision, radar, lidar) to overcome individual sensor limitations
• Improve machine learning algorithms for moving object detection and tracking

Safety features:

• Standardize redundant systems and integrate parachute compatibility
• Enhance obstacle detection for challenging targets like strings and wires through improved algorithms and potentially complementary sensor technologies
• Design frangible components that meet FAA energy transfer requirements
• Develop sensor performance monitoring that alerts pilots to degraded obstacle avoidance conditions

Operational guidance:

• Provide clear, honest documentation about obstacle avoidance system capabilities and limitations
• Include specific guidance on environmental conditions that affect sensor performance
• Offer training materials that help operators understand when to rely on automation versus manual control
• Document appropriate use cases and Part 107 category eligibility

Data sharing:

• Collaborate with regulators and researchers to refine injury risk models with real-world incident data
• Share anonymized near-miss data to improve industry-wide understanding of operational challenges

Both Skydio and DJI have demonstrated commitment to safety through continuous product improvement. As the industry leader in consumer/prosumer drones, DJI’s extensive platform range offers agencies mission-appropriate options across weight classes. Skydio’s focus on AI autonomy and U.S. manufacturing addresses specific first responder needs around hands-free operation and Blue sUAS requirements. Continued innovation from all manufacturers—including advances in sensor technology and obstacle detection algorithms—will expand the safety-capability frontier for all operators.

Conclusion: A Path Forward for Responsible Drone Deployment

Drones have become indispensable tools for modern first responders, providing situational awareness that saves lives and improves operational effectiveness. The physics of mass and kinetic energy is straightforward: heavier drones carry greater harm potential in failure scenarios. However, this physical reality must be contextualized within the broader operational picture—actual failure rates, safety system effectiveness and limitations, mission requirements, and alternative risk profiles, including lacerations and airborne hazards.

The solution is not prohibition of capable platforms but rather risk-proportionate operations informed by realistic understanding of technology capabilities:

• Match aircraft selection to genuine mission needs
• Optimize flight parameters to minimize exposure
• Understand and plan for the limitations of obstacle avoidance technology
• Implement multiple layers of technical and procedural safeguards aligned with FAA guidance
• Maintain rigorous training that includes both technology capabilities and constraints
• Foster transparent communication with communities, emphasizing privacy protections

The near-miss incident serves not as evidence that advanced drones are inherently too dangerous for public safety work, but rather as an educational example of the real-world challenges that all vision-based obstacle avoidance systems face. Moving objects, thin obstacles, weather conditions, and challenging lighting situations represent detection scenarios that challenge current sensor technology across the entire industry. Understanding these limitations allows operators to compensate through intelligent flight planning, environmental assessment, and appropriate use of manual control when conditions warrant.

By combining physics-informed risk assessment, proven safety technologies, realistic appreciation of current technological limitations, professional operational protocols aligned with Part 107 Subpart D, and open dialogue with communities, first responders can confidently deploy drones in service of public safety—protecting both the communities they serve and the officers who serve them.

As obstacle avoidance technology continues to evolve—potentially incorporating radar, lidar, or other complementary sensors—the safety margins will improve. Until then, the human operator remains the most critical safety component, armed with training, judgment, and understanding of both what the technology can do and what it cannot.

References

1. FAA Part 107 Subpart D – Operations Over People – 14 CFR Part 107
2. Advisory Circular 107-2A – Small Unmanned Aircraft Systems (sUAS) – FAA AC 107-2A
3. FAA ASSURE UAS Collision Severity Reports – FAA Safety Research
4. “Terminal Velocity and Impact Characteristics of Small UAS” – AIAA Journal
5. “Drone-human collision tolerance and injury risk assessment” – Science Direct
6. “A dataset for assessing the safety of small unmanned aerial vehicles during human encounters” – Nature Scientific Data
7. “Safety Risk Analysis for sUAS Operations” – PMC Review
8. NYPD Drone as First Responder Program – Skydio News, Sept 23, 2024
9. MITRE Drone as First Responder Resources – Public Safety Toolkit
10. “No Kings, No Parachutes, No Problem?! NYPD Flies Skydio Drones Over Crowds” – DroneXL, Oct 18, 2025

---

Disclaimer: This article has been reviewed by several drone industry specialists on accuracy and relevance. It presents general safety considerations for drone operations based on publicly available research and regulatory guidance. It should not be construed as definitive operational guidance, legal advice, or manufacturer-specific recommendations. Agencies must consult FAA regulations (particularly Part 107 Subpart D and AC 107-2A), legal counsel, manufacturer specifications, and conduct site-specific risk assessments when developing drone programs. Impact energy estimates are based on published research and simplified models; actual collision outcomes depend on numerous variables including impact angle, body location, protective equipment, structural deformation, laceration potential, and individual physiology. Failure rate statistics vary by platform, operating environment, and mission profile. Obstacle avoidance system performance varies with environmental conditions, and all vision-based systems have inherent limitations with moving objects, thin obstacles, and challenging lighting conditions regardless of manufacturer.