The　advent　of　autonomous　vehicles　(AVs)　with　self-parking　capability　puts　forward　new　requirements　for　parking　policy　and　management.　In　a　transportation　system　with　both　general　and　dedicated　AV　lanes,　this　work　aims　to　examine　and　compare　the　effectiveness　of　different　parking　policies　on　the　morning　commute.　To　be　specific,　three　policies　of　parking　space　allocation,　namely,　first-come-first-service　(FCFS),　advanced　reservation　(AR),　and　their　combination　(CB),　are　proposed　to　guide　the　commuters＇　travel　choices.　Commuters　can　drive　regular　vehicles　(RVs)　alone　on　the　general　lane　or　ridesharing　shared　autonomous　vehicles　(SAVs)　on　the　dedicated　AV　lane.　The　occurrence　conditions　and　the　resultant　SAV　penetration　rates　for　various　scenarios　under　three　parking　policies　are　revealed.　The　area　division　for　various　scenarios　with　respect　to　two　key　parameters　is　presented,　and　the　typical　departure　patterns　are　plotted　as　well.　The　numerical　results　clarify　that　the　CB　policy　always　has　a　better　effect　on　the　total　travel　cost　than　FCFS　and　AR　policies.　The　optimal　CB　policy　is　quantified　in　terms　of　the　optimal　amount　of　all　parking　spaces　and　the　proportion　of　unreserved　parking　spaces.　These　findings　help　to　improve　the　spatial　and　temporal　distribution　of　traffic　flow　and　allocate　transportation　resources　efficiently　in　the　coming　era　of　autonomous　driving.