DSP-controlled boost converter program flow

[zilog]

The main workings of this smps is a boost-converter without current feedback, but with ADC-connections for both input (Vdc) and output (Vo) voltages. The DSP can command any duty cycle (DC) based on programmatic commands. The input can range between 8-16V (car battery) and the output should ideally be kept fixed at 17V.

I am thinking in the direction of hysteretic control, where I try to keep the boost-converter discontinuous most of the time (the steady state current drawn will not demand continuous mode). I also dont need any overcurrent protection for the output as I trust the circuits I will be powering. The only scary thing I still haven't completely figured out is what to do with the duty cycle during power-on.

My thoughts:

I will program the ADC to sample both Vdc and Vo at the last 20% T, when the inductor normally would be dry given my dead time margins, and always use a duty cycle that lets the inductor reach a preset peak current (Ipk) which is 0.7A in my case, based on the input voltage.
During power-up Vo will be one diode drop less than Vdc, and the inductor current will go continuous very fast and saturate/break the switch. Since I dont use any current feedback, I have no way of knowing when this happens. I am thinking in the way of as long as Vo is below say (Vdc - 1), to start at 0.1% DC, and measure the delta in Vo for each switch cycle. As long as I dont have at least say 0.05V delta for each cycle, I increase the DC in 0.1% steps. When I have passed 0.05V delta, I keep the DC constant, if it falls below I increase it again (the delta will be smaller the higher Vo is since a constant amount of energy is transferred each cycle).
As soon as I have come above say 15V, I command the DC that will give the target Ipk for each cycle, and switch to the hysteretic control scheme where pulses are skipped as long as the voltage is high enough.

Will this do, under the presumption that the output will never be loaded with a large enough step to cause the inductor current to go continuous after a successful start-up? Is there any good way of calculating what Ipk the inductor will se given steady state and different sizes of the control steps?