The design of a GNSS-based high integrity navigation system for automotive applications presents many challenges that are related to the complicated propagation channel encountered in a typical road environment. The presence of obstacles and/or reflectors on or beside the road such as large trucks, buildings, trees, walls, overpasses or tunnels can result in a large variation in the number of received signals, and a large variation regarding the quality of the GNSS measurements and the independence of these errors both in time and across measurements. As an example, the well documented occurrence of Non Line-of-Sight (NLoS) tracking conditions will tend to create a GNSS pseudorange measurement error distribution with a heavy positive tail. This same situation might, if the obstacle at the origin of the NLOS situation is large enough to affect several satellites, result in correlating the error of multiple measurements. These phenomena are not necessarily well handled by traditional integrity mechanisms because they create breaches of fundamental assumptions (for instance ability to perform CDF overbounding or the assumption of measurement independence). This might result in an under-estimation of the targeted rate of Hazardous Misleading Information (HMI) required by the application. u-blox is actively working on various integrity solutions that would allow to be resistant to the above phenomenon. These solutions exploit differently the characteristics of the GNSS measurement errors, typically through the use of a strict measurement selection process and their exploitation by either a sequential filter (to capitalize on time filtering) or an advanced snapshot filter (taking advantage of the fine modeling of the GNSS measurement errors). Of course, the combined effect of strict measurement selection and sky-view obstruction can create frequent situations where a GNSS-only navigation system with integrity becomes unavailable. In this case, it appears critical to be able to use additional and complementary sensors. The typical sensors available on cars are IMUs and odometers, and these sensors are often used to improve accuracy and availability. However, their use does not necessarily conform to stringent integrity requirements, particularly ensuring that the required HMI rate is maintained. The objective of this paper is to present how the IMU and WT measurements are integrated in both the sequential and snapshot integrity solutions. Both methods require distinct solutions that allow for different advantages and pose constraints. The key element is to ensure that the solution remains appropriately bounded in various situations that range from the availability of a large number of good GNSS measurements to a reasonably long GNSS outage. For a sequential integrity algorithm, sensor data fits naturally into the estimation framework. The major challenges are in ensuring that the sensor models (both deterministic and stochastic) are adequate to maintain valid bounds, and that the non-linearity in propagation is handled properly. For the snapshot integrity algorithm, there is no built-in notion of bound propagation, and this concept must be explicitly handled by a new algorithm. When bound propagation is used to reduce latency or computational load, one option is to use GNSS delta-phase observations. During GNSS outages, this is clearly not possible, and IMU and WT sensors are the only available option. In this case, the propagation algorithm is somewhat similar to that used in the EKF, though it requires special care. Results presenting the bounding capability of both the sequential and snapshot solutions will be shown on an extensive set of more than 100 hours of real data collected in a large number of environments going from city centers to open sky roads. The performance in terms of protection level magnitude as a function of the “GNSS environment” will be presented as well as the comparison between both integrity solutions. Conclusions and way forward will then be formulated.