UAV Energy Consumption Models for Wireless Systems Research: Model Selection and Misconceptions
summary
The gist
Uncrewed aerial vehicles (UAVs) are gaining increasing attention in wireless systems, providing new opportunities to expand the reach and improve the quality of wireless servicesDespite their…
In short
The paper reviews energy consumption models for Uncrewed Aerial Vehicles (UAVs) relevant to wireless systems research. It categorizes UAV types and flight phases, critiques common model misuse, and highlights significant errors that arise from selecting inappropriate energy models. The goal is to provide practical guidance for choosing accurate analytical models for deployment and trajectory planning.
Key concepts
- UAV Classification
- UAVs are grouped by airframe design into Fixed-wing (FWUAVs), Rotary-wing (Single/Multi-rotor), and Hybrid UAVs. These classifications dictate how energy is used during different flight phases, such as takeoff, cruise, and loiter. Understanding this classification is crucial because the energy models used must match the specific physical characteristics of the aircraft being modeled.
- Flight Phases
- UAV operations are broken down into distinct phases: takeoff, ascending, level flight (cruise), and loiter. Each phase consumes energy differently; for example, cruise is usually the most efficient. The paper emphasizes that using a model designed for one phase when the UAV is in another can lead to substantial errors in calculating total energy use.
- Analytical Models
- These are physics-based models used to estimate energy consumption. The authors focus on these because they are tractable and adaptable for optimization problems. They deliberately exclude overly complex, data-driven models, prioritizing practical usability for deployment and trajectory planning in wireless research.
Terminology used across episodes
This episode discusses
- UAV Energy Consumption Models for Wireless Systems Research: Model Selection and Misconceptions · Paper Radio
The paper
UAV Energy Consumption Models for Wireless Systems Research: Model Selection and Misconceptions · Read on arXiv
University of California, Irvine
Transcript
Introduction to the show: ident: Electrical Engineering. The day in electrical engineering: signals, systems and the hardware that carries them.
Rosa: I'm Rosa.
Dev: And I'm Dev, here with Taro.
Rosa: We're looking at the paper: "UAV Energy Consumption Models for Wireless Systems Research".
Dev: Here's the gist of the paper. Uncrewed aerial vehicles (UAVs) are gaining increasing attention in wireless systems, providing new opportunities to expand the reach and improve the quality of wireless servicesDespite their versatility, UAVs are limited by available energy onboard, which results in significant challenges in deploying UAV-enabled wireless systemsModeling energy consumption is an essential component of the deployment and trajectory optimization of UAVsThis article presents a comprehensive overview of UAV energy consumption models, with a focus on their relevance to wireless systems researchWe deliberately exclude data-driven and overly complex models to provide clear and practical guidelines for their use in wireless systems researchWe begin by categorizing the most common types of UAVs and describing the typical flight phases considered in the literatureWe then review existing energy consumption models, focusing on their scope with respect to UAV types and flight phasesWe also discuss common mistakes in the use of these models and highlight the existing gaps in the literatureIn particular, we show how the use of a wrong model can lead to significant errors in energy consumption calculationsFinally, we emphasize the need to develop energy consumption models for missing scenarios UAV Classification and Operational Phases The paper categorizes UAVs based on airframe characteristics and describes typical flight phases considered in wireless systems researchFigure 2 shows the specifications of a typical midsize battery-powered UAVFixed wing UAVs use rigid, non-flapping wings to generate lift and remain airborne through forward motion, similar to conventional airplanesFixed wing UAVs are known for their high endurance and their ability to cover long distances due to their aerodynamic and energy efficiencyA typical flight profile of a FWUAV begins with the takeoff phase, where the UAV starts from a stationary position on the ground and ends when the UAV is airborne, i.e., when the generated lift is sufficient to achieve and maintain levitationOnce in the air, FWUAVs enter the ascending phase, during which they gain altitude by maintaining a forward motion with a positive pitch angleAfter reaching the desired altitude, FWUAVs switch to the level flight phase, where they maintain a constant altitudeIn this phase, the speed can be variable, but for a constant speed, the flight phase is called the cruise, the most energy-efficient phase of flightFWUAVs may also perform 3D maneuvers such as turning, changing altitude, or adjusting their courseIn the context of wireless systems, the loiter phase is particularly relevant for FWUAVsLoiter phase is particularly relevant to FWUAVs in the context of wireless systems researchHigh-Altitude Platform Stations (HAPS) differ significantly from typical FWUAVsHAPS operate at much higher altitudes, often above 12 km, with endurances ranging from weeks to monthsRotary-wing UAVs are generally classified into two subclasses: single-rotor UAVs (SRUAVs) and multi-rotor UAVs (MRUAVs)Single-Rotor UAVs can carry moderate payloads, operate at low to moderate altitudes, and cruise at moderate speedsMulti-rotor UAVs typically carry smaller payloads, operate at lower altitudes, cruise at slower speeds, and exhibit shorter endurance due to higher energy consumption and less efficient aerodynamic designHybrid UAVs combine the advantages of both FWUAVs and MRUAVs, offering VTOL capability alongside the aerodynamic efficiency of fixed-wing flight Review of Energy Consumption Models The paper reviews existing energy consumption models and presents their corresponding scenarios for which they are applicableUAV energy consumption models can be broadly categorized into analytical (physics-based) models, empirical or experimentally derived models, and data-driven approachesAnalytical models offer a tractable and adaptable framework that can be easily incorporated into optimization problemsWe deliberately exclude overly complex formulations that depend on detailed parameters or are data-driven, and instead focus on models with practical usability for the purpose of deployment and trajectory planning in wireless systemsTable I provides a summary of the available energy consumption models for wireless systems researchThe objective is not to exhaustively review UAV energy consumption models, but to provide practical guidance for the selection of models tailored to wireless research applications Common Misconceptions and Model Errors The authors highlight common misconceptions regarding the use of existing models and identify scenarios where appropriate energy consumption models do not existA common misuse is to treat the takeoff of SRUAVs and MRUAVs to different altitudes in the same way, thereby ignoring significant differences in energy consumptionEnergy consumption models for the ascending and descending phases of fixed-wing UAVs and HUAVs have not been exploredA common mistake is to approximate these phases using piecewise level flight phasesApproximating level flight phase with a series of piecewise cruise flights introduces a noticeable errorThe models in Table I are used for these calculationsIn general, this approximation is only valid when altitude variations are small and vertical speed remains nearly constantA common mistake is to use SRUAV hover energy consumption models instead of those for MRUAVs, despite differences in aerodynamic efficiency that can lead to significant estimation errors Impact of Flight Phase Assumptions The comparison of flight ranges across different UAV types demonstrates the importance of selecting an energy consumption model that matches the specific UAV typeUsing a SRUAV energy consumption model instead of a MRUAV energy consumption model can overestimate the range in the level phase by up to 100%Even a small climb rate leads to approximately 5% to 50% error in the estimation of the flight range, depending on the UAV configurationApproximating level flight phase with a series of piecewise cruise flights introduces a noticeable errorThis approximation becomes increasingly inaccurate when the required acceleration is highIn free flight, where both altitude and speed vary, the resulting error is even more pronouncedThe models in Table I are used for these calculationsApproximating level flight phase with a series of piecewise cruise flights introduces a noticeable errorInaccuracy increases with larger changes in altitude and velocity, particularly when UAVs experience significant vertical accelerationThe models in Table I are used for these calculations Future Research Directions The article concludes by emphasizing the need to develop energy consumption models for missing scenarios and suggests future research directionsFuture research may focus on developing lightweight, physics-based energy consumption models for different UAV types and flight phases that remain insufficiently explored for wireless systems research purposesThis work has been submitted to IEEE for possible publication Acknowledgments This work was supported in part by the NSF Award ECCS-2513060 References
1: M.
Rosa: Mozaffari, W.
Dev: Saad, M.
Rosa: Bennis, Y.-H.
Title and authors: Rosa: So we're looking at this paper now called "UAV Energy Consumption Models for Wireless Systems Research: Model Selection and Misconceptions" by Barzegaran and Jafarkhani. It’s about how we model the energy use of uncrewed aerial vehicles, which is crucial because battery life really dictates how far those wireless services can reach.
Dev: Exactly. This paper aims to give us a clear picture of the models out there, but it focuses on selecting the right one for our actual deployment and trajectory planning. It deliberately skips the overly complex or data-driven stuff to keep things practical for research applications <ref:2605.14116#pg2>.
Taro: So, what does this paper actually do for us when we're trying to fly these drones? What’s the main takeaway from looking at these energy models?
Rosa: Well, the authors start by sorting out the different types of UAVs we see in wireless research. They categorize them into things like Fixed-wing UAVs, High-Altitude Platform Stations, and those rotary-wing ones like single-rotor and multi-rotor platforms <ref:2605.14116#pg2>.
Dev: And they tie those types to typical flight phases, which are things like takeoff or level flight. The whole point is to show how these specific UAV types interact with those specific flight phases when we try to model energy consumption <ref:2605.14116#pg2>.
Taro: I’m thinking about how this helps me test failure modes in the real world. If we use the wrong model, are we going to misrepresent how much energy is actually needed for a maneuver?
Rosa: That’s a big concern. The paper points out common mistakes people make when using these models, and it highlights scenarios where the right model just doesn't exist yet <ref:2605.14116#pg2>.
Dev: It specifically calls out treating the takeoff of single-rotor and multi-rotor UAVs as if they use the same energy consumption model for different altitudes. That's a common misuse that leads to significant errors <ref:2605.14116#pg2>.
Taro: So, if I’m designing a mission profile, I need to be careful about that altitude dependency during takeoff?
Rosa: Precisely. They show how the energy consumption for ascending and descending phases of fixed-wing UAVs and Hybrid UAVs hasn't been fully explored yet <ref:2605.14116#pg3>.
Dev: And they warn against approximating the level flight phase with a series of piecewise cruise flights, which introduces noticeable errors when acceleration gets high or when altitude changes are significant <ref:2605.14116#pg3>.
Title and authors: Taro: So, for real-time control loops, that means we can't just use a simple approximation for sustained flight if the dynamics change?
Rosa: That’s right. The models they present in Table I are used for these calculations <ref:2605.14116#pg3>.
Dev: They also show how using a single-rotor hover model instead of a multi-rotor one can cause significant estimation errors because the aerodynamic efficiency is different <ref:2605.14116#pg3>.
Taro: That makes sense, because if we’re controlling a platform, those small differences in efficiency translate directly into battery life predictions that could be off by a lot.
Rosa: The main goal here is to provide practical guidelines for choosing the model that fits the specific UAV type and flight phase we are working with in wireless systems research <ref:2605.14116#pg2>.
Dev: It’s about moving away from just having a bunch of models and instead having a systematic way to select the correct one, which is currently lacking in that literature <ref:2605.14116#pg2>.
Taro: So, if we want better trajectory optimization for dynamic maneuvers, what’s the next step suggested by this paper?
Rosa: The authors emphasize that there's a real need to develop energy consumption models for those missing scenarios <ref:2605.14116#pg2>. They suggest future research should focus on developing lightweight, physics-based models for flight phases that are currently underserved in wireless systems work <ref:2605.14116#pg2>.
Dev: From an engineering standpoint, it means we need to start thinking about how to integrate model selection logic directly into our AI frameworks so they check if the chosen model is valid for the specific UAV and phase before making a decision Rosa.
Taro: So, it’s not just about choosing a model; it’s about building a system that knows *which* model to use based on where we are in the flight profile?
Rosa: Exactly. It prevents us from getting misleading design conclusions by applying models outside their intended scope <ref:2605.14116#pg2>.
Dev: And for line-of-sight maintenance, it suggests we need to be very careful about distinguishing between single-rotor and multi-rotor hover energy consumption because using the wrong one can lead to significant estimation errors <ref:2605.14116#pg3>.
Taro: So, when I’m planning a loiter pattern for a fixed-wing UAV, I need to be acutely aware that the model I pull for cruise might not be accurate if the altitude changes are large <ref:2605.14116#pg3>.
Rosa: That’s right. And we have to acknowledge that even small climb rates can lead to errors in range estimation of about five percent to fifty percent, depending on the UAV configuration <ref:2605.14116#pg3>.
Title and authors: Dev: That variability is what keeps me up at night for real-time loops. If the energy model has a fifty percent error margin, that affects our entire safety case <ref:2605.14116#pg3>.
Taro: It really shows that the impact of flight phase assumptions isn't small; it can be huge when you’re looking at range or endurance predictions <ref:2605.14116#pg3>.
Rosa: So, to wrap up this paper, "UAV Energy Consumption Models for Wireless Systems Research: Model Selection and Misconceptions" gives us a framework to be more disciplined about our modeling choices.
Dev: It’s about avoiding those common pitfalls where we use an inappropriate model because the correct one just doesn't exist for a specific UAV type or phase <ref:2605.14116#pg2>.
Taro: The implication is that future work needs to focus on developing those physics-based models for the missing scenarios, like those ascending and descending phases of fixed-wing UAVs <ref:2605.14116#pg3>.
Rosa: We need to move toward having model selection built into the AI planning framework so it knows exactly which energy consumption model is valid for the current situation <ref:2605.14116#pg2>.
Dev: It’s about making sure that when we talk about trajectory optimization, we’re using models that actually reflect the vertical speed and acceleration components of the flight path <ref:2605.14116#pg3>.
Taro: For me, it means when I test recovery strategies for a drone, I need to be hyper-aware of how much the model might overestimate or underestimate energy during those rapid changes in vertical motion <ref:2605.14116#pg3>.
Rosa: So, this paper is a call to action for the research community to stop using models blindly and start being very specific about what UAV we are modeling and what phase of flight we are in <ref:2605.14116#pg2>.
Dev: It’s a practical guide for making deployment strategies that are truly feasible within the onboard battery limits Rosa. We need to get those energy calculations right before any AI system starts planning real missions.
Taro: It’s about ensuring that our autonomy systems can handle the actual energy costs of hovering versus loitering correctly, so we don't design a mission that looks good on paper but drains the battery in practice <ref:2605.14116#pg3>.
Rosa: We’ve covered the UAV types, the flight phases, and where these models fail right now. That’s what this paper gives us for today regarding UAV energy consumption models for wireless systems research <ref:2605.14116#pg2>.
The paper's summary: Rosa: So, this paper is essentially giving us a guide on which energy models to use when we’re modeling drone battery life for wireless systems <ref:2605.14116#pg2>.
Dev: Right, it’s not about finding one perfect model; it's about knowing when to use which one based on the UAV type and the flight phase <ref:2605.14116#pg2>.
Rosa: They break down the UAV types first, like fixed-wing versus rotary-wing drones, and then they map those types onto specific flight phases, such as takeoff or cruise <ref:2605.14116#pg2>.
Dev: And they look at the existing models for each of those scenarios to see what actually applies <ref:2605.14116#pg3>.
Rosa: The main thing they point out is that using the wrong model can lead to really big errors in energy calculations, which messes up the entire deployment strategy <ref:2605.14116#pg3>.
Dev: They highlight a few common mistakes we see, like treating the takeoff energy for single-rotor and multi-rotor drones as if they are similar, which is just not true <ref:2605.14116#pg3>.
Rosa: And they show that approximations, like using a series of piecewise cruise flights to model level flight for fixed-wing drones, introduce errors when acceleration gets high <ref:2605.14116#pg3>.
Dev: It’s a practical warning because those errors can be as big as fifty percent in the range estimation if you don't account for vertical speed and acceleration correctly <ref:2605.14116#pg3>.
Rosa: So, the implication is that we need to stop using models blindly and start checking if they’re actually right for what we’re simulating <ref:2605.14116#pg2>.
Dev: That means when we build our trajectory optimization systems, they have to incorporate a check—a model selection logic—to ensure the energy consumption model matches the specific drone and phase <ref:2605.14116#pg2>.
Rosa: They also stress that there are whole scenarios where no good model exists yet, so we need to develop those for future research <ref:2605.14116#pg2>.
Dev: That means the next step for us is figuring out how to build those new, physics-based models for the ascending and descending phases of fixed-wing drones and hybrid drones <ref:2605.14116#pg3>.
Rosa: It’s about making sure our AI systems select deployment strategies that are actually feasible within the battery limits, not just theoretical ones <ref:2605.14116#pg2>.
Dev: This is crucial for real-time loops because we need to know the safety case when we're planning maneuvers that require high acceleration <ref:2605.14116#pg3>.
Rosa: So, this paper gives us a framework for being more disciplined about our modeling choices in wireless research <ref:2605.14116#pg2>.
Dev: And it forces us to think about how the energy models affect things like line-of-sight maintenance between single and multi-rotor drones <ref:2605.14116#pg3>.
Rosa: So, we’ve covered the paper's main summary and the big warning about model selection errors. Next up, we look at how this affects our actual mission planning for deployment outside the lab.
The paper's improvements: Rosa: So, we’re looking at the suggested improvements in this paper now, which is basically how we can fix those model selection problems we talked about earlier <ref:2605.14116#pg2>.
Dev: The authors aren't just pointing out mistakes; they are actually proposing a new way to build the systems that use these models <ref:2605.14116#pg2>.
Rosa: They suggest implementing model selection logic directly into the AI framework so it checks if the chosen energy model is valid for the specific UAV type and flight phase before it makes a decision <ref:2605.14116#pg2>.
Dev: That’s smart because it prevents us from making misleading design conclusions by applying models beyond their intended scope, which is something we see happening all the time in real-time loops <ref:2605.14116#pg2>.
Rosa: And they also propose developing new analytical models for those missing scenarios, like the ascending and descending phases of fixed-wing UAVs and hybrid drones <ref:2605.14116#pg3>.
Dev: That means we need to start building those specific energy consumption models instead of relying on piecewise approximations that introduce errors when acceleration is high <ref:2605.14116#pg3>.
Rosa: They also emphasize refining how we model the level flight phase by moving away from that series of cruise flights and toward a more accurate representation of free flight <ref:2605.14116#pg3>.
Dev: That’s where the physics-based models come in, trying to accurately capture both vertical speed and acceleration components, which is what drives those errors up when things get dynamic <ref:2605.14116#pg3>.
Rosa: So, the big picture here is that we need to move toward having model selection built into the AI planning framework so it knows exactly which energy consumption model is valid for the current situation <ref:2605.14116#pg2>.
Dev: And this directly impacts our safety case because we can't trust a trajectory if the underlying energy calculation is flawed <ref:2605.14116#pg3>.
Rosa: It’s about ensuring that when we talk about trajectory optimization, we’re using models that actually reflect the real physics of the flight path <ref:2605.14116#pg3>.
Dev: Plus, they stress developing those missing models for scenarios where no good model exists yet, which is a big gap in the current literature <ref:2605.14116#pg2>.
Rosa: So, the implication is that future work needs to focus on developing lightweight, physics-based models for flight phases that are currently underserved in wireless systems research <ref:2605.14116#pg2>.
Dev: We need to get those energy calculations right before any AI system starts planning real missions, and this paper gives us a clear roadmap on how to do that Rosa.
Rosa: It’s about making sure our autonomy systems can handle the actual energy costs of hovering versus loitering correctly, so we don't design a mission that looks good on paper but drains the battery in practice <ref:2605.14116#pg3>.
Dev: That means when we talk about real-time control loops, we have to be hyper-aware of how much the model might overestimate or underestimate energy during rapid changes in vertical motion <ref:2605.14116#pg3>.
Rosa: So, this paper is a call to action for the research community to stop using models blindly and start being very specific about what UAV we are modeling and what phase of flight we are in <ref:2605.14116#pg2>.
Conclusion: Rosa: So we’re wrapping up this discussion on "UAV Energy Consumption Models for Wireless Systems Research: Model Selection and Misconceptions" <ref:2605.14116#pg2>.
Dev: Essentially, the paper shows us that picking the wrong energy model can lead to really big errors in our deployment calculations <ref:2605.14116#pg3>.
Rosa: They’re pushing for a framework where AI systems explicitly check if the chosen model is actually valid for the specific UAV type and flight phase they are simulating <ref:2605.14116#pg2>.
Dev: That means we need to stop using models blindly and start being very specific about what drone we’re modeling and exactly which phase of flight it’s in <ref:2605.14116#pg3>.
Rosa: They also highlight the need for new physics-based models for scenarios where no good model exists yet, like the ascending and descending parts of fixed-wing drones <ref:2605.14116#pg3>.
Dev: That’s a big ask because it changes how we design the real-time loops to ensure the safety case holds when things get dynamic <ref:2605.14116#pg3>.
Rosa: So, for someone listening, this means being disciplined about our modeling choices so we don't end up with impractical solutions based on flawed assumptions <ref:2605.14116#pg2>.
Dev: It’s about ensuring that when we talk about trajectory optimization, the energy calculations actually reflect the vertical speed and acceleration components of the flight path <ref:2605.14116#pg3>.
Rosa: We’ve covered how to spot common model errors and where more research is needed for missing scenarios in this paper <ref:2605.14116#pg2>.
Dev: The implication is that future work has to focus on developing those lightweight models for different UAV types and flight phases that are currently insufficiently explored <ref:2605.14116#pg2>.
Rosa: It’s a call to action for the research community to be more specific about what they are modeling in wireless systems research <ref:2605.14116#pg2>.
Dev: We need model selection built into the AI planning framework so it knows exactly which energy consumption model is valid for the current situation <ref:2605.14116#pg2>.
Rosa: It’s about making sure our autonomy systems can handle the actual energy costs of hovering versus loitering correctly, not just what looks good on paper <ref:2605.14116#pg3>.
Dev: I think we need to keep pushing on those real-world failure modes, especially when it comes to how a system recovers gracefully from an energy estimation error <ref:2605.14116#pg3>.
Rosa: I agree, and that’s what this paper sets up for the next phase of research in this area <ref:2605.14116#pg2>.
More episodes
- Daily Summary for 2026-11-19
- 2111.07910-Mask-guided Spectral-wise Transformer for Efficient Hyperspectral Image Reconstruction
- 2203.02149-HDNet: High-resolution Dual-domain Learning for Spectral Compressive Imaging
- 2205.10102-Degradation-Aware Unfolding Half-Shuffle Transformer for Spectral Compressive Imaging
- 2311.10959-Structure-Aware Sparse-View X-ray 3D Reconstruction
- 2403.04116-Radiative Gaussian Splatting for Efficient X-ray Novel View Synthesis
- 2410.02572-Combining Pre- and Post-Demosaicking Noise Removal for RAW Video
- 2604.17453-Learned Nonlocal Feature Matching and Filtering for RAW Image Denoising
- 2603.20387-End-to-End Multi-Task Learning for Adjustable Joint Noise Reduction and Hearing Loss Compensation
- 2507.20825-Chirp-Permuted AFDM: A Versatile and Multifunctional Waveform for 6G