Pollutant emissions standards (such like EURO 6 in Europe) are increasingly severe and force a search of new in-cylinder strategies and/or aftertreatment devices / schemes at a reasonable cost. On a conventional Diesel engine an excess of air is usually used to allow very high combustion efficiencies and reasonable levels of soot which can then be after-treated in a diesel particulates filter (DPF). As a consequence, NOx emissions cannot be easily after-treated (lean NOx trap (LNT) and selective catalytic reduction (SCR) are quite expensive even if effective, solutions), as a result they are generally controlled by means of internal measures such as High Pressure (HP) or Low Pressure (LP) exhaust gas recirculation (EGR). In light of ever more stringent NOx emissions regulations, NOx aftertreatment devices seem to be becoming unavoidable.The goal of the paper is to test stoichiometric combustion on an automotive Diesel engine (with the aim of reducing NOx emission with simple 3-way catalyst) while trying to overcome the traditional drawbacks of this combustion mode (large increase of Particulate Matter (PM), CO and unburned HC emissions and noticeable decrease of engine global efficiency).An experimental study is carried out on a 2.0L automotive common-rail High Speed Direct Injection (HSDI) Diesel engine, equipped with a variable geometry turbocharger and an intercooler. Two air loop strategies to get stoichiometric conditions are investigated: low temperature throttled fresh air and high temperature EGR. In each case, various injection strategies are tested, especially while varying the timing of the injection splits and the fuel mass distribution. A particular focus is made on in-cylinder pollutant emissions production (PM, NOx), combustion efficiency, and thermal efficiency. Characteristics and advantages of each air intake strategies are studied and compared. A heat release analysis helps to understand the various combustion modes, in particular to determine the proportion of fuel which burns in premixed conditions. Finally, the most promising paths towards stoichiometric Diesel combustion are underlined. For instance, it seems that when stoichiometry is obtained by intake air with EGR, the combustion processes is more efficient (about 93%). In turn, the throttling of the intake air allows the most advantageous level of PM emissions (soot emission below 2 FSN was achieved). For the best cases, stoichiometric diesel combustion sacrifices specific fuel consumption by around 6%. As a conclusion, some evolutions on engine design are proposed to further improve the concept.