Q.Why are curved roads banked?

Banking tilts the road so the horizontal component of the normal reaction, $N\sin\theta$, itself supplies (or contributes to) the centripetal force, reducing dependence on the limited and unreliable force of friction.

On a FLAT (unbanked) curved road, the entire centripetal force needed for a vehicle to turn must come from static friction between the tyres and the road, $f_s=\frac{mv^2}{r}$. This is a poor long-term solution for two reasons: friction has a hard upper limit ($f_{s,max}=\mu_sN$), which caps the maximum safe turning speed, and in real conditions $\mu_s$ is far from constant -- it drops sharply on a wet, oily, dusty, or otherwise non-uniform road surface, making the safe speed limit unpredictable and potentially dangerously low exactly when it matters most (e.g. in rain). By tilting the road surface at an angle $\theta$ (banking), the normal reaction N -- which is always present regardless of friction -- now has a HORIZONTAL component $N\sin\theta$ that points towards the centre of the curve and can, by itself (at the 'most safe speed' $v_s=\sqrt{rg\tan\theta}$), supply the ENTIRE centripetal force needed, with zero reliance on friction at all. For other speeds in a realistic range around $v_s$, friction still contributes, but only up to its safe (well below limiting) capacity, giving the road a much wider genuinely safe speed range ($v_{min}$ to $v_{max}$) than an unbanked road ever could.

Compare the flat-road case (all centripetal force from limited, unreliable friction) with the banked-road case (the tilted normal reaction's horizontal component supplies some or all of it).

Thinking banking eliminates the NEED for friction entirely, rather than reducing the DEPENDENCE on it -- friction is still needed (and used) at speeds other than the one exact 'safest speed' $v_s$.