Estimating power consumption of different type of radio access technology

A method comprising:

• determining types of radio access technologies supported by an apparatus, wherein the apparatus supports at least two types of radio access technologies;
• providing inputs to machine learning models, wherein the machine learning models comprise a corresponding machine learning model to each type of radio access technology, and for each type of radio access technology parameters associated with that type of radio access technology are provided as an input to the corresponding machine learning model;
• obtaining, for each type of the radio access technologies, a power consumption estimate provided by the corresponding machine learning model;
• and
• determining a total cost of error function based on comparing a power consumption measurement obtained for the apparatus and a sum of the power consumption estimates obtained for each type of radio access technologies.

A non-transitory computer readable medium comprising program instructions for causing an apparatus to perform at least the following:

• determine types of radio access technologies supported by the apparatus, wherein the apparatus supports at least two types of radio access technologies;
• provide inputs to machine learning models, wherein the machine learning models comprise a corresponding machine learning model to each type of radio access technology, and for each type of radio access technology parameters associated with that type of radio access technology are provided as an input to the corresponding machine learning model;
• obtain, for each type of the radio access technologies, a power consumption estimate provided by the corresponding machine learning model;
• and
• determine a total cost of error function based on comparing a power consumption measurement obtained for the apparatus and a sum of the power consumption estimates obtained for each type of radio access technologies.