Symphony: Simple Phase Control for Wave Energy Systems
summary
The gist
Symphony is a novel nonoptimisation-based (NOB) controller inspired by phase control techniques for wave energy devices that achieves close-to-optimal performance across evaluated sea states…
In short
Symphony is a novel non-optimization-based controller for wave energy devices inspired by phase control. It uses a hierarchical structure with an upper loop for feedforward control and a lower loop for tracking. This approach achieves close-to-optimal performance across various sea states while guaranteeing hard motion constraints, even with measurement noise and model uncertainties.
Key concepts
- Hierarchical Structure
- The controller is organized into two main parts: an upper loop that generates a feedforward force and a velocity reference, and a lower loop that tracks this reference. This separation allows for better control of different aspects of the system's motion.
- Implicit Gaussian Differential Equation (IGDE)
- This mathematical equation is used to generate desired trajectories for the WEC velocity. Solving it ensures that the resulting trajectories stay within a predefined safe set, which guarantees that physical motion constraints (like maximum position and velocity) are always met.
- Filtering Sliding-Mode Control (SM)
- The lower loop uses this control method to track the desired velocity reference. It is chosen for its robustness against unmodelled disturbances and measurement noise, effectively filtering out noise while maintaining stable tracking performance.
Terminology used across episodes
This episode discusses
The paper
Symphony: Simple Phase Control for Wave Energy Systems · Read on arXiv
Pedro O. Fornaro, Eugenio M. Gelos, Demian Garcia-Violini, John V. Ringwood
Centre for Ocean Energy Research at National University of Ireland, Maynooth · Consejo Nacional de Investigaciones Científicas y Tecnicas at Universidad Nacional de Quilmes, Argentina
Despite significant advancements in controlling wave energy converters, efficiently achieving real-time energy-maximising control remains a challenge. This paper addresses this obstacle by introducing Symphony, a novel non-optimisation-based (NOB) controller inspired by phase control techniques for wave energy devices. Symphony is a suboptimal control that inherits all the advantages of NOB controllers, such as simplicity of design, implementation, tuning, and use, while additionally providing hard-motion-constrained reference trajectories and close-to-optimal performance across the evaluated sea states, even in the presence of measurement noise and model uncertainty. Symphony is designed using a hierarchical structure: By relying on an implicit Gaussian differential equation that requires an excitation force estimate, an upper loop provides a close-to-optimal motion-constrained velocity trajectory and a feedforward control term. Then, a user-flexible lower loop tracks the desired velocity trajectory by combining feedback and feedforward terms. In this paper, Symphony performance is assessed in terms of power absorption. Statistically representative results are obtained using multiple sea-state realisations across varying peak wave periods, considering measurement noise and radiation model uncertainty. The results presented in this paper demonstrate the potential of Symphony, which achieves, in most of the evaluated scenarios, more than 90% of the power absorption obtained by a spectral (optimisation-based) control solution.
DOI: 10.1109/TCST.2026.3733146
Transcript
Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.
Rosa: We're looking at the paper: "Symphony: Simple Phase Control for Wave Energy Systems".
Dev: Here's the gist of the paper. Symphony is a novel nonoptimisation-based (NOB) controller inspired by phase control techniques for wave energy devices that achieves close-to-optimal performance across evaluated sea states, even in the presence of measurement noise and model uncertainty <ref:2608.09525#pg2> Controller Architecture Symphony is designed using a hierarchical structure consisting of an upper loop and a lower loop <ref:2608.09525#pg4>.
Rosa: The upper loop provides a feedforward control force and a suboptimal velocity reference that is proportional to an excitation force estimate, effectively phase-matching the WEC velocity with the wave excitation force estimate <ref:2608.09525#pg4>.
Dev: Then, a user-flexible lower loop tracks the desired velocity trajectory by combining feedback and feedforward terms <ref:2608.09525#pg4>.
Rosa: Forward-invariant trajectories are obtained by solving an implicit Gaussian differential equation (IGDE) <ref:2608.09525#pg5>.
Title and authors: Rosa: So we're looking at this paper called "Symphony: Simple Phase Control for Wave Energy Systems." It’s about a new way to control these wave energy devices.
Dev: Yeah, it tackles the challenge of getting real-time energy-maximizing control without needing heavy optimization. This paper introduces Symphony as a novel nonoptimisation based controller inspired by phase control techniques used in wave energy devices.
Taro: So, what’s the main idea behind this approach? Is it simpler to implement than those complex optimization based methods we see in some of the other papers?
Rosa: Exactly, that’s a big part of it. Symphony is designed to have simplicity in design, implementation, tuning, and use while still providing hard motion-constrained reference trajectories and close-to-optimal performance across different sea states even when there's measurement noise or model uncertainty.
Dev: That sounds like it’s aiming for that sweet spot between complexity and reliability. The paper sets up a hierarchical structure for this controller, which is pretty interesting because it separates the tasks into an upper loop and a lower loop.
Taro: A hierarchical structure means there's some kind of layering to how the control works, right? What does that layering actually look like in practice when you’re trying to keep something stable?
Rosa: Well, the upper loop provides a feedforward control force and a suboptimal velocity reference. This reference is proportional to an excitation force estimate, which essentially helps phase-match the wave energy converter's velocity with that estimated excitation force.
Dev: And then the lower loop tracks that desired velocity trajectory by combining both feedback and feedforward terms, which is where the actual tracking happens. But how do they get those trajectories in the first place?
Taro: That’s where this paper introduces solving an implicit Gaussian differential equation, or IGDE. That means they use that equation to generate forward invariant trajectories, which are solutions that stay inside a safe set, regardless of measurement noise or model uncertainty.
Rosa: So what does that mean for the real world? It means these solutions are guaranteed to comply with hard motion constraints for both position and velocity even if things get noisy or the model is slightly off.
Dev: That’s a significant improvement over some other methods where you might need constant re-tuning when uncertainty pops up. They handle those uncertainties by using an excitation force estimate, which is what you need, and they recommend using an estimator from a prior study to handle the noise in the upper loop.
Taro: So if we think about a real deployment on a device out at sea, how does this guarantee that it won't violate physical limits? Does it stop it from hitting something or exceeding its safe operational bounds?
Title and authors: Rosa: It guarantees that the solutions generated by solving that IGDE stay inside what they define as a forward invariant set, which is basically their safe set. This ensures compliance with hard motion constraints for both position and velocity, even when dealing with measurement noise and model uncertainty <ref:2608.09525#pg2>.
Dev: I’m also interested in how they handle the reference generation without needing a full forecast. They get the motion-constrained velocity reference from the upper loop without resorting to forecasting, only needing a current wave excitation force estimate.
Taro: That sounds like it avoids one of the biggest headaches in control systems, which is when you need perfect future predictions that are impossible to get accurately for dynamic waves. It relies on what’s happening right now.
Rosa: Right, and they use a numerical approximation obtained by controlling a virtual wave energy converter to synthesize that velocity reference, which helps bridge the gap between the theoretical model and what the actual system is doing.
Dev: That brings us to how they actually implement the tracking in the lower loop. They combine feedforward and feedback terms for tracking, which is standard stuff, but they use a filtering sliding mode control, or SM control from twenty-one, as their tracking controller <ref:2608.09525#pg2>.
Taro: Why sliding mode? It’s known for being robust to unmodelled disturbances and measurement noise, right? How much robustness are we talking about when you combine that with the other elements of Symphony?
Rosa: The filtering sliding mode control is chosen specifically because it has this property where the filtering structure helps reject measurement noise and considerably reduces the effects of that noise in the control action <ref:2608.09525#pg4>.
Dev: And when you look at the performance evaluation, they show that Symphony achieves more than ninety percent of the power absorption you get from a spectral optimisation based control solution in most of the tested scenarios. That’s a pretty high number for a nonoptimisation based controller.
Taro: So it’s not just about being simple; it’s actually competitive in terms of performance, but they are still comparing it to those full optimisation methods. What's the trade-off there?
Rosa: The main point is that Symphony is the first nonoptimisation based controller we know of that can provide a velocity reference which respects motion constraints, and this happens independently of WEC model errors and measurement noise <ref:2608.09525#pg4>.
Dev: And when it comes to radiation errors, they’re quite robust. They show less than four percent variation in power absorption for the S3 case with a parametric error in the radiation model <ref:2608.09525#pg9>. That's less than five percent deviation.
Title and authors: Taro: Less than five percent variation is pretty tight when you’re dealing with physical components that might degrade or change over time in the real world. How does that compare to the effort they are using?
Rosa: They show that the effort required by the sliding mode control is four orders of magnitude smaller than the feedforward control term, even across different scenarios with varying radiation errors <ref:2608.09525#pg10>.
Dev: And there’s a security margin called Delta z that they mention, which you can select depending on how robust the tracking controller is. That gives the design flexibility for deployment conditions.
Taro: So, to sum up what we’ve seen from this paper on Symphony: it's a method that uses phase control inspiration to get a hierarchical structure with an IGDE for trajectory generation, which keeps you within safe sets, and then combines it with filtered sliding mode control for robust tracking.
Rosa: And the result is that it gets you close to optimal power absorption—over ninety percent in most cases—while maintaining constraint satisfaction even when things are noisy or uncertain <ref:2608.09525#pg4>.
Dev: It’s a strong position for a real-time system, especially since it avoids the heavy computational load of full optimisation, and they claim this is the first nonoptimisation based controller to achieve close-to-optimal performance in constrained scenarios with minimum recalibration <ref:2608.09525#pg4>.
Taro: So for someone listening who just wants to know what this means for the field, it suggests that you don't necessarily need the most complex optimization framework to get very good results in wave energy control; you can achieve close-to-optimal performance with a simpler design and less reliance on perfect models.
Rosa: Right, so we’ve looked at the Symphony paper today. It shows how using phase control techniques can lead to a robust controller that handles uncertainties better than some alternatives.
Dev: It really highlights the benefit of combining different ideas—the IGDE for trajectory generation and the filtering sliding mode control for tracking—to satisfy those hard motion constraints <ref:2608.09525#pg4>.
Taro: I just wonder about that hard force constraint mechanism they mention in their conclusion, since they suggest designing one next. That feels like the next logical step to move this from a promising paper to something you can actually deploy reliably <ref:2608.09525#pg2>.
Rosa: Definitely. It shows the path forward for making these controllers truly ready for real-world deployment where those hard limits are non-negotiable.
Dev: Yeah, we'll keep an eye on that future work, because getting those constraints right is key to making sure this controller actually works in practice.
The paper's summary: Rosa: So, to wrap up this part, Symphony is essentially this new way to control wave energy devices by drawing inspiration from phase control methods for waves, aiming for simple design but high performance across different sea states and even with noise and model errors.
Dev: Right, so we're talking about this hierarchical structure again—the upper loop handles the feedforward force and a reference velocity based on an excitation force estimate, while the lower loop tracks that using feedback plus feedforward terms.
Taro: And the key move here is solving that implicit Gaussian differential equation to generate trajectory references that stay inside a safe set, which means they respect hard position and velocity constraints even when things get messy.
Rosa: That’s what I mean—it’s about getting those motion-constrained trajectories without needing perfect forecasting or constantly recalibrating the model every time the wave conditions shift.
Dev: And it uses a filtering sliding mode control for that tracking, which is good because it has that structure to reject measurement noise pretty effectively.
Taro: I think what’s really interesting is how they manage the radiation errors; they show less than four percent variation in power absorption even when there's a parametric error in the radiation model.
Rosa: That level of robustness, especially across different sea states, is what makes this approach so promising for real-time operation on a device out at sea.
Dev: And the performance metrics are strong; they’re getting over ninety percent of what you'd expect from those heavy optimization based controllers in most scenarios.
Taro: So, it suggests that we can get very good results without needing the full computational burden of spectral methods, which is a big deal for deployment.
Rosa: Exactly. It’s about finding that balance between simplicity and getting close to optimal performance under real-world constraints.
Dev: Now, the paper also points out some areas for future work; they suggest designing a hard force constraint mechanism as a next step to make it even more reliable in practice.
Taro: That makes sense because while they handle motion constraints well, explicitly defining those hard force limits sounds like the next logical hurdle for making this truly deployable.
The paper's improvements: Tom: So, we're moving on to the specific improvements they propose for Symphony after showing off how it works in principle, and I think those refinements are where the real practical value lies.
Rosa: The paper suggests focusing on a hard force constraint mechanism as a next step because while they handle motion constraints well, explicitly defining those hard force limits sounds like what you need to make this truly ready for deployment.
Dev: That makes sense, because having those explicit boundaries written down helps build that safety case we talked about earlier when the world misbehaves.
Taro: I agree with Rosa; it moves it from a theoretical success in simulation to something that’s guaranteed safe when the real-world forces push things past the expected limits.
Rosa: It’s about taking that general safety set idea and making it a strict rule, which is important for field roboticists like myself who need those hard stop conditions.
Dev: And from a controls standpoint, defining that hard force limit gives us something concrete to work with instead of relying solely on the soft constraints inherent in the Gaussian modulating envelope.
Taro: That gives the system more predictable recovery modes when things get really extreme, like when a wave hits unexpectedly hard.
Rosa: So, they’re essentially tightening up the safety net by adding those explicit force limits on top of what they already have for position and velocity.
Dev: It’s about ensuring that even if the model gets wildly wrong, you still have a baseline physical protection mechanism kicking in without needing a massive recalibration effort.
Taro: And I think that robustness against extreme events is what will really make this system useful for autonomy in dynamic environments where things aren't perfectly known.
Rosa: It sounds like the next phase of development for Symphony isn't just tweaking the math, but grounding it firmly in physical limits that a device has to obey.
Dev: So, if we can nail that force constraint mechanism, we move from a very robust controller to one that’s truly ready for long-term operation under unpredictable sea conditions.
Conclusion: Rosa: So, to wrap up this whole discussion on Symphony: Simple Phase Control for Wave Energy Systems, we're seeing a controller that balances simplicity in design with high performance in real-world conditions while staying under hard motion constraints.
Dev: Yeah, it’s a nonoptimisation based approach that manages to get over ninety percent of the power absorption from those more complex optimization methods across most evaluated sea states.
Taro: For me, what stands out is how little variation in power absorption you see when you introduce radiation model errors; less than four percent for the S3 case is pretty solid for something operating outside a lab.
Rosa: It means this isn't just theoretical stuff; it has shown robustness when the world misbehaves, which is exactly what we need when we put these devices out in the field.
Dev: The paper also points to the future work needing a hard force constraint mechanism, which is crucial for moving this from a promising concept to something that’s deployable without constant manual adjustments.
Taro: I think that focus on those explicit force limits is where we need to go next because it gives us a clearer way to define when the system should just stop or change its strategy under extreme load.
Rosa: Exactly, so while Symphony gets you there with the IGDE and the filtering sliding mode control, defining those hard force limits will be the final piece of making it truly ready for unpredictable operation.
Dev: It’s a solid piece of work on how to get close to optimal performance without needing massive computational power for full optimization.
Taro: Anyway, this paper shows that phase control ideas can lead to a very robust controller when you're dealing with uncertainty in wave energy systems.
More episodes
- 2608.13438-ContactGuard: Pre-Contact Execution Monitoring with Action-Conditioned Latent World Models
- 2610.34593-Future Prediction as Training-Only Video Supervision for Pretrained VLA Policies
- 2610.35385-EmbodiedLab: Beyond Manipulation, Benchmarking Embodied Agents That Act to Know
- 2610.36378-Data-Driven Hamiltonian for Neural Optimal Control
- 2610.37275-A Heuristic for Scheduling Autonomous Vehicles at Intersections
- 2610.37288-Dynamic Coordination of Autonomous Vehicles at Intersections
- 2610.37632-Real-Time Coordination of Autonomous Vehicles at Intersections
- 2611.13779-Softness is something you also need! Why compliance helps control: information metrics of a passive walker with compliant ankle
- 2611.13411-Residual-Corrected Physics-Informed Neural Network KKL Observers with Input-to-State Stability Guarantees
- 2611.14488-Is Your Robot's Memory Just a Clock? Temporal Shortcuts in Memory-Augmented VLA Policies