A decrease of worldwide on-road emissions of vehicles is demanded over the next few years and compliance with current regulations needs achievement. This invention aims at reducing the fuel consumption and tailpipe emissions of hybrid electric vehicles (HEVs) by implementing a real-time energy management system capable of optimally adapting to the driving style of the current user.
Patent status
SUBMITTED
Priority Number
102020000008686
Priority Date
22/04/2020
License
INTERNATIONAL
Market
The percentage of hybrid electric vehicles (HEVs) produced and sold by car manufacturers has been growing significantly in recent years, especially in the passenger car sector. The proposed invention, designed for the energy management of HEVs, is aimed at a market that extends on a global scale, mainly to automotive manufacturers (OEMs) and their first level suppliers (TIER1s). However, in the initial stages of product marketing, the geographical segmentation variable will be privileged, converging towards the European market.
Problem
From a sustainable mobility perspective, hybrid electric vehicles are presented as one of the solutions that can lead to fuel savings and therefore tailpipe emissions. However, due to their complexity, they require the definition of energy management systems that can be implemented in on board electronic control units. The goal of these systems is to ideally achieve near-optimal fuel savings on different driving scenarios and styles, limiting the computational effort compatibly with the technological limits currently present. Among on-board energy management systems, rule-based ones are the most widely used solution thanks to their ease of implementation, but they show a degradation of performance in terms of fuel savings as driving scenarios and styles vary. In this context, the presented solution aims to define an energy management system that can be implemented on board capable of overcoming the limits of the solutions currently used. This solution would lead to multiple socio-economic benefits. In fact, it could not only support car manufacturers in complying with the emission limits imposed by future regulations but also lead to significant advantages in terms of fuel savings, thus making the invention also attractive from the point of view of the end user.
Current technologies limits / Solutions
At the current state of the art, few technologies have been proposed to adapt the optimal control of a hybrid electric powertrain to the driver’s driving style. For example, patent US9361272B2 proposes a feedback system to the driver about the ideal position of the accelerator pedal aiming at fuel saving. However, this system requires the user’s willingness to adapt to the pre-calibrated vehicle control logic and allows for improvements only in a short time horizon. Alternatively, the US0070679A1 patent proposes a driver model based on Machine Learning capable of predicting the vehicle speed profile related to the driver in a limited time horizon for the consequent improvement of energy management through ECMS logic. However, the proposed system is not capable of optimizing the operation of the HEV powertrain over the entire driving mission. Alternatively, the patent US7954579B2 proposes to use an artificial intelligence (AI) agent as the control logic of a hybrid electric powertrain. Training of the AI agent is achieved through a supervised learning based on the optimal HEV control policies extracted from an off-line global optimizer. However, the applicability of the considered invention is demonstrated only in homologation driving cycles, while it does not allow adaptation to the driving style of a specific user in real-world scenarios.
Killer Application
The main applications of this technology are:
Technology and our solution
The technology represents an innovative system for the real-time optimization of the on-board energy management of hybrid electric vehicles. The main target of such a system is to minimize the fuel consumption as well as the tailpipe emissions while being adaptive with respect to the driving style of the vehicle user.
The system would include an advanced control unit for hybrid powertrains that can edit the implemented control algorithm as a function of the specific driver before starting a new journey. Additionally, the system would include an interface (e.g. haptic, visual) through which the driver can identify in the vehicle control system, and a storage system (either physical or virtual) for the Artificial Intelligence agent representing the HEV tailored control algorithm.
Once the driver is identified by means of the interface, the system extracts the tailored HEV control algorithm from the storage system and loads it into the advanced HEV supervisory control unit, thus allowing the subsequent optimization of the HEV powertrain operation during the journey.
The potential product could find application on a wide variety of HEVs (e.g. passenger cars, light-duty or heavy-duty vehicles). Among these, preliminary results have proved the capability of the technology of reducing by 20% the fuel consumption produced by a light-duty commercial vehicle.
Advantages
Compared to the current state-of-the-art solutions, the proposed technology shows the capability of effectively adapting the optimal control of hybrid electric powertrains to the driving style of a specific user. Such a condition is obtained through an Artificial Intelligence agent which learns the optimal control policies in an off-line phase for different personal driving scenarios.
In addition, the optimal reduction in fuel consumption and tailpipe emissions is achieved over the entire journey, and the user can maintain their usual driving feeling.
Roadmap
The first hardware-in-the-loop prototype of the technology will be realized in the next few months through a Proof of Concept financed project of Politecnico di Torino. AVL Italia Srl has shown its willingness to support the activity through the provision of methodologies and tools necessary for validation, as well as through support for the identification of market interests related to the proposed technology.
Once the results of the prototype massive testing will be available, the best solution for finalizing the technology on the market will be exploited.
TRL
The team
