Among the different opportunities to save fuel and reduce CO2 emissions from internal combustion engines, great attention has been done on the waste heat recovery: the energy wasted is, in fact, almost two thirds of the energy input and even a partial recovery into mechanical energy is promising. Usually, thermal energy recovery has been referred to a direct heat recovery (furtherly expanding the gases expelled by the engine thanks to their high pressure and temperature) or an indirect one (using the thermal energy of the exhaust gases - or of any other thermal streams - as upper source of a conversion power unit, which favors a thermodynamic cycle of a suitable working organic fluid). Limiting the attention to the exhaust gases, a novel opportunity can be represented by directly exploiting the residual pressure and temperature of the flue gases through an Inverted Brayton cycle (IBC), in which the gases are expanded at a pressure below the environmental one, cooled down and then recompressed to the environmental pressure. Considering the thermodynamic conditions of the exhaust gases, expansion and compression must be done into dynamic machines, so making profit of the technology of the turbocharging group. The useful power of an IBC-based recovery unit is strictly related to the behavior of the machines chosen as IBC turbine and compressor (pressure ratio vs. mass flow rate and efficiencies), considering that they run at the same speed. Therefore, an experimentally based mathematical model has been developed to evaluate the coupling of an IBC-based recovery unit with a real turbocharged diesel engine whose real behavior was measured on a dynamic test bed. In particular, experimental data of the engine have been used as boundary conditions of the direct recovery group, considered as a bottomed recovery unit. In this way, the room of recovery of the unit has been assessed evaluating also the possibility to accept a higher engine backpressure in order to have a higher recoverable power, optimizing the whole system. An overall net CO2 saved of about 7 g/kWh has been achieved for a specific working point.